The detection of network intrusion by evolutionary immune algorithm with clonal selection

In the paper the application of artificial immune systems as a heuristic detection method for algorithmic information security incidents of intrusion detection systems is proposed. A theory of clonal selection has been selected from the existing computational models of artificial immune systems with the necessary features for building adaptive intrusion detection systems. The efficiency of clonal selection algorithm is increased forming high affinity detectors) by modification of clonal selection algorithm by applying an external structure optimization, which principle is based on the application of the evolutionary algorithm strategy. For affinity calculation in work the metrics “coordination percent” is used. Additionally, in the paper the pseudorandom numbers generator on the basis of Blum-Blum-Shub’s algorithm is applied. The empirical results of the evolutionary immune clonal selection algorithm effectiveness have been received by testing on a set of test data according to the procedure of research. Comparative analysis with similar efficiency of the developed evolutionary immune clonal selection algorithm, constructed on the other methods of artificial intelligence, is performed. Conclusions on the efficiency of application of the evolutionary immune clonal selection algorithm selection at the solution of a problem of deliberate changes detection on a controlled data are formulated according to the results of the research.