¡¡	Chinese Journal of Computers   Full Text
Title	An Algorithm of r-Adjustable Negative Selection Algorithm and Its Simulation Analysis
Authors	ZHANG Heng1) WU Li-Fa2) ZHANG Yu-Sen2) ZENG Qing-Kai3),4)
Address	1)(Institute of Communication Engineering, PLA University of Science and Technology, Nanjing 210007) 2)(Institute of Command Automation, PLA University of Science and Technology, Nanjing 210007)
Year	2005
Issue	No.10(1614¡ª1619)
Abstract &   Background	The Artificial Immune System (AIS) is introduced briefly, then as one of the main algorithms in AIS, negative selection algorithm is discussed. In order to reduce the number of holes which is inevitable in negative selection algorithm, an algorithm of r-adjustable negative selection algorithm is put forward. The new algorithm can reduce the number of holes through adjusting threshold of matching. The simulation results show that both the number of iteration to generate mature detector and the number of holes decline quickly, while the rate of detecting abnormity raises. keywords artificial immune system; negative selection algorithm; threshold of matching; hole background The paper is supported by the National High Technology Research and Development Program of China (863 Program) under grant Noª±2002AA141090(Research and Development on Security Key Technology of Server), Noª±2004AA147070(¡°Research and Development on Security Evaluation of System Platforms¡±) and the National Natural Science Foundation of China under grant Noª±60473053(¡°Study on Semantic Constraint Approach to Control Program¡¯s Behavior¡±). The security of program roots from two ultimate problems: One is the behavior of program cannot abide by the intention of its designer, another is the access control of program cannot abide by the least privilege principle. In those research projects, static and dynamic methods, confining and monitoring technology are combined to turn the protection of program from passive to active. During the research process, immune mechanism is applied to the modeling and control of the program behavior, and also applied to the automated modeling and test of security function. In recent years, the main work concerned with security that finished by authors¡¯ research team list as follow: developed an applied distributed network intrusion detection system ¡°NetNumen¡± which based on rule and can combine anomaly detection with misuse detection, developed an applied military security operation system which can reach B security level. Now the research team focuses on security test and embed operation system. The immune mechanism is used to solve the key problem of program security which is security state monitor and get nice results. But one of the important algorithms in immune mechanism named negative selection algorithm exists hole which can lead to False Negatives. In this paper, the authors applied a simple way through adjusting the matching threshold to reduce the number of holes. The simulation which adopts data got from experiment shows that both the number of iteration to generate mature detector and the number of holes decline quickly, while the rate of detecting abnormity raises.