An Agent-Based Simulation Approach For Criminal Network Analysis

As the struggle of anti-terrorism and anti-crime go fierily,digital anti-terrorism, anti-crime technology has been researched widely.These technologies include using statistical approach to analysis crime data;using data mining classify crime data and find useful association rules;and using social network analysis method to capture the insight structure of terrorist organizations.The security bureaus of different countries pay very attention on crime data collection,database mining and related database construction.The intelligent information technology such as Data mining and social network analysis is a hot topic in anti-crime and anti-terrorism security bureaus of different countries.Social Network Analysis(SNA) is a promising method to analyze crime network structures.A social network is just a network that has a member of nodes with the connection between members to express their relations.Analyzing criminal network by SNA method have a certain outcome,it can find the important nodes in criminal networks and point out potential structures.However,social network analysis is a point in time in the static analysis of criminal networks of nodes and structures.Criminal may adjust when facing damage in some cases.Social network analysis could not handle the characteristics of dynamic change of criminal networks.Because of the limitation of traditional social network analysis methods,we consider using Agent technology.Agent is an independent software entity in special environment.It can interact with the environment by flexible and intelligent way,learn from experience and communicate with other Agent or process.Multi-Agent cooperation can effectively solve large and complex problems.Agent-based simulation technology can simulate the evolution of the military organization with network structure,command controlling organization and social organization.In this paper,we look forward to finding the useful strategies to destabilize the criminal organization by using the multi-Agent system to simulate the characteristic of dynamic changes of criminal organizations combined with the analysis methods of individual members of the network and the overall structure of the network.We collect criminal data to construct numbers of virtual criminal networks,and do serial experiments to find the strategies to destabilize the criminat organization based on them.The experimental results show that the Agent-based simulation approach for criminal networks analysis caneffectively find the useful strategies to destabilize the criminal organization.We can extend more helpful methods for anti-crime and anti-terrorism based on this framework.