Font Size: a A A

Artificial Immune Principle And Its Application In Intrusion Detection

Posted on:2006-12-22Degree:MasterType:Thesis
Country:ChinaCandidate:J LiFull Text:PDF
GTID:2208360152997176Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Many computational intelligence research fields such as artificial neural network,genetic algorithm have been applied to the solution of computing problem. Theexplorer for other computational intelligence technique, for instance, artificialimmune system, swarm intelligence, etc, appear since last years and many problemscouldn't get solution. Nature immune system is a very complex system and manyimmune mechanisms are unfamiliar and unrecognized. Most of the artificial immunesystem researches simulate little part of the nature immune system. Manyresearchers have suggested intrusion detection system based on immune theory bystudying artificial immune system.After the introduction of nature immune, artificial immune, intrusion detectionand theirs development, main feature this dissertation studies several researchdirection of the artificial immune system and its applications.First of all, dissertation studies the representation of the antibody and antigenthat have been seldom studied. This dissertation suggests that wavelet analysis canreduce feature population numbers effectively and extend the covered space ofproblem field by studying the application of mathematic method in representationform of antibody and antigen.Stimulated by the resemble of human immune system and the intrusion detectionsystem, a immune network applied in distributed intrusion detection system issuggested.Selecting a good fitness function is important when applying the geneticalgorithm in artificial immune system. There is little research result in this aspect forartificial immune theory has focused on the framework of artificial immune systemlast years. In fact, the selection of fitness function is difficult by the research resultof this dissertation. This dissertation suggests several fitness functions that issuitable for the intrusion detection system.Finally, a intrusion detection system model based on major histocompatibilitycomplex and fuzzy logic decision algorithm is suggested. This model has two par.
Keywords/Search Tags:artificial immune, genetic algorithm, major histocompatibility complexity, fitness function, antibody, antigen, immune network, fuzzy logic, intrusion detection
PDF Full Text Request
Related items