Research And Application Of Improved Clone Selection Algorithm In Computer Virus Detection

Driven by scientific and technological empowerment and economic benefits,computer viruses are becoming more and more intelligent and frequent.Unknown viruses and virus variants also appear at any time,which brings great challenges to computer virus detection.Artificial immune system is based on biological immune system,biological immune system naturally has the ability to distinguish between "self" and "non-self",which is very suitable for application in the field of computer anti-virus.In order to solve the problem that unknown viruses and virus variants cannot be detected by the static detection method,a clonal selection algorithm combining forgetting mechanism and targeted update strategy is proposed and applied to computer virus detection.The main work of this thesis is as follows.(1)Study the static detection method for computer viruses,analyze the reasons for its inability to detect unknown viruses and virus variants,and propose solutions.The computer virus static detection method relies on the computer virus signature database,which is composed of known virus characteristics.The problem of information lag makes it impossible to detect unknown viruses and virus variants.The clonal selection algorithm carried out directional proliferation and mutation of the detector,which had the ability to detect unknown viruses and virus variants.Therefore,the clonal selection algorithm was selected to carry out the research.(2)In view of the long training time of clonal selection algorithm and insufficient diversity of antibodies,attenuation theory and interference theory in the forgetting mechanism were introduced to perform selective forgetting of intermediate repetitive antibodies,so as to shorten the training time of the model and accelerate the updating speed of antibodies.Aiming at the blindness and uncertainty caused by random updating of clonal selection algorithm,a new updating strategy was proposed to guide antibodies to update in the direction of global optimization,so as to improve the optimization accuracy and convergence stability of the algorithm.The comparison experiment with CEC test function proves that the improved algorithm has better optimization accuracy,convergence speed and convergence stability.(3)In view of computer virus feature inventory information lag,update is not timely,virus variant easily lead to the failure of the feature code and other problems.The improved clone selection algorithm is used to train computer virus samples,and a representative computer virus signature database is constructed.The similarity matching method is used instead of the one-by-one comparison method to obtain approximate results,so as to detect unknown viruses and virus variants.Henchiri and CILPKU08 data sets were used to complete different comparative experiments.The experiments proved that the improved virus detection model has better detection effect and is an effective computer virus detection model.Finally,the model is applied to computer virus detection system and the system test is completed.