A New Immune Model For Computer Viruses

Computer viruses and worms are an increasing problem throughout the world because of their new evolution trends includes growing birth rate and very rapid spreading speed. Traditional detection techniques are failed to defeat such computer viruses effectively. An emergent problem today is to find a real-time strategy to defeat computer viruses, especially new viruses. The Artificial Immune System(AIS) which stimulates the Biology Immune System(BIS) is an investigative hotspot because of its potential capability to solve this problem. Many researchers, such as S.Forrest, S.Homeyr, Kephart, are involved into this field and bring up a series of Computer Virus Immune Systems(CVIS). However, some limitations,such as too huge self universe and low generating efficiency of mature detectors,make it hard to put current CVIS into practice. In order to fetch up these limitations,a new immune model for computer virus inspired by biology immune system is proposed in this paper. Quantitative description of the model is given. The problem of dynamic evolution of self and nonself are solved which reduces the size of self universe effectively. Moreover, a new method for immature detector generation is proposed by using evolutive gene library. It improves the generating efficiency of mature detector and keeps false-negative and false-positive at a reasonable level. Specifically, the contribution of the dissertation includes: establishing a dynamic and quantitative model; giving a dynamic model of self to slove the problem of the dynamic representation of self in a real-time environment; proposing a new method for immature detector generation by using evolutive gene library. It improves the generation efficiency of mature detector and keeps false-negative rate and false-positive rate at a reasonable level.