Study On Behaviour Of Complicated Network System Forecast Based On Catastrophe Theory

With the development and the popularization of Internet, the information age has come to us. Now it has become much easier to be attacked through the Internet, therefore doing research on behaviour of a complicated network system, especially on the alert of computer attack becomes very important. It is not enough to satisfy the user by studying how to protect the computer after an attack had happened, it is more beneficial for the researcher to do the research while the computer is under attack than when it is over. We can provide the largest benefit to users to avoid any loss on the part of the user.The thesis firstly introduces the academic significance of the subject, points out the latest research result of Catastrophe Theory both in domestic and overseas, analyzes the current situation of the development of the Catastrophe Theory and network security and demonstrates the question settled by the thesis and scheme. Secondly, it tells us the problem of Catastrophe Theory and complicated system in detail including the general situation of Catastrophe Theory, odd point and stability of structure, the object of Catastrophe Theory, the application of Catastrophe Theory, Zeeman catastrophe structure, catastrophe features, type of catastrophe, trait of complicated system and so on. Because it is very important to apply Catastrophe Theory to the information security field, the next chapter also tells us in detail about how to build catastrophe model of the unusual behaviour of network and how to estimate and test model parameter. Upon comparison with other theses this thesis finds it essential to put forward the cooperation of Catastrophe Theory and Fuzzy Math, the cooperation of Catastrophe Theory and Neural Network and the cooperation of Catastrophe Theory and IDS based on network intrusionalert model. Among these the thesis puts forward how to apply RBF (Radial Basis Function) and WNN (Wavelet Neural Network) network to forecast unusual behaviour point of network. Compared with various kinds of forecast models of unusual behaviour point of network, the thesis puts forword the improving forecast model of ARMA in order to ensure the rate of accuracy.The paper applies Catastrophe Theory into field of security alert of information. It extends the field of application of Catastrophe Theory and furthers its application in other fields.