Research On Clone Selection Algorithm Based On Quantum Theory In IDS

At present, the network security situation is becoming increasingly severe. As a kind of passive defense tool, the firewall still has some shortages. Intrusion Detection System is receiving increasing attention. For the threat to organism posed by virus is similar with the threat to network system posed by network intrusion, it has become the focus of the study that how to maintain network security in the way that organism used against virus.IDS is a kind of network security products, the detection rate is an important indicator of good or bad. If we want to improve the detection rate under the condition of limited storage space, it is necessary to optimize the detector, the good ones is preserved, while the bad ones is deleted. To achieve this goal, lots of Immune algorithm is developed, such as Negative selection algorithm, GA, Clone selection algorithm. Clone selection algorithm is a kind of process of evolution. GA breeds the next generation through hybrid, while the Clone selection algorithm does this through cloning, at the same time it Improve the affinity of the individual through controllable variation in order to achieve the evolution of populations. In order to overcome the question of detector redundancy and the low speed of evolution, a IDS server-intrusion detection engine system is raised. Besides of this, two jobs is accomplished. First, the detector structure in the intrusion detection system is improved. It is based on the polymorphism of quantum code, and increases the amount of information taking advantage of quantum storage systems, besides, the needed measurements is saved in binary storage. Second, the quantum-inspired clone selection is raised through improving the traditional clone selection algorithm. In this algorithm, the quantum compute technology is combined with traditional clone selection. Detector is optimized taking advantage of selection and variation. Only the best measurement result is saved so that the redundancy is avoided. In the course of variation, the direction of variation of offspring detector is guided by the known of the current optimal detector. In this way, the speed of evolution is raised while the problem of prematurity is avoided. Last, the algorithm is tested and the result shows that the algorithm can improve detection rate and reduce omitted rates effectively...