Research On The Algorithms Of Network Intrusion-detection Based On Swarm Intelligence Computing

Now, network tech. is developing in high speed and information is exploding, so that we must face the problem that keeping the informations safely. It is impossible to make software with 0 bug, for our thinking's localization and software engineering is not perfect. As a researchable focus in information safty field, intrusion-detection has won many attentions. At first,this paper discuss the correlative knowledges on intrusion-detection, and sort the researchable methods and analysis the features of each method. Secondly, we discuss the principle of the swarm intelligence computation based on the frame of competition and collaboration and it is proved that differential evolution(DE) with keeping best individual is convergent by Markfov-chain and artificial fish swarm(AFS) is proved convergent by cotheorem of compress mapping. In addition, this paper research three key technologies including feature selection,clustering and association rules mining of intrusion-detection, and design the algorithm by DE and AFS. At last, it is proved by experiments that using swarm intelligence algorithm to solve the three key problems above is effective.