Research On The Model Of Quantum-inspired Immune Intrusion Detection System

With the rapid development of computer network technology and the widely use of many information systems in recent years, the security of network is becoming increasingly prominent. As an active defense technology, the intrusion detection plays an important role in the study of computer security technology.Artificial immune system imitates the function of natural immune system, which can identify known and unknown antigen, and protect itself successfully. Artificial immune system can detect normal and abnormal behavior patterns effectively and make up the disadvantage of intrusion detection system on accuracy and efficiency. Intrusion detection algorithm and model based on artificial immune system is a focus on the research field of intrusion detection recently.By introducing the current research on the artificial immune intrusion detection, we analyze the mechanism of artificial immune system and study the immune-based intrusion detection system from detector tolerance modules and intrusion detection module in this thesis. The purpose of this research is to design a network intrusion detection model using artificial immune and principles of quantum computing, and improve the deficiency of network intrusion detection system.The main content of this thesis is as follows:(1) Firstly, for the deficiency of two traditional intrusion detection which proposed by Forrest and Kim respectively, a new intrusion detection model based on quantum theory and immune principles is proposed in this thesis. The architectural framework, work principles, primary features and the deficiencies of the new intrusion detection model is introduced simultaneously.(2) Secondly, Study on the algorithm of the detector generating module and intrusion detection module which are the two main modules of the quantum-inspired immune intrusion detection model. The basic idea of the detector generating algorithm is to train a high diversity of antibody through self-learning mode firstly, and then to use it to the real data. For the shortcomings of r-continuous-bit matching rule, an improved r-continuous-bit matching rule is designed in this thesis. With the new matching rule, the intrusion detection algorithm has more efficient detection rate.(3) Finally, the usability of the model proposed in this thesis is examined by using digital simulation technology. All the training and testing datasets are based on the intrusion evaluation data sets. The simulation results show that the new algorithm is more efficient than other methods.