A detector evolution algorithm based on Immunization Strategy of Complex Networks

The Detector Generation Algorithm based on Niching Strategy (DGANS) could pass good genes to the next generation and maintain the diversity of the population. In this paper, a Detector Evolutionary Algorithm based on Immunization Strategy of Complex Networks (DEAISCN) is proposed, in which the immunization strategy of complex networks is studied to improve the evolutionary process of mature detectors in DGANS. The Affinity Function is used to optimize the choice of parent detectors in DEAISCN, then it is not necessary to acquire the characteristic information of all the individuals, besides the non-self modes are more likely to be selected. Simulation results show that DEAISCN has lower missing alarm rate and false alarm rate than DGANS, and DEAISCN outperforms DGANS as the encoding length increases.