CAJ | 학술논문

针对传统的否定选择算法和V‐detector算法存在的“黑洞”以及检测器冗余问题，论文提出了优化算法。在检测器生成阶段，候选检测器与自体数据和成熟检测器同时进行耐受训练，比较结果确定检测器半径，减少检测器的重复覆盖；在检测器成熟阶段，计算自体重复覆盖率剔除完全被覆盖的检测器，克服了“黑洞“及冗余问题。最后，通过Iris数据比较了三种算法。实验结果显示，改进算法的检测器数量大幅减少，利于实践应用。
침대전통적부정선택산법화V‐detector산법존재적“흑동”이급검측기용여문제，논문제출료우화산법。재검측기생성계단，후선검측기여자체수거화성숙검측기동시진행내수훈련，비교결과학정검측기반경，감소검측기적중복복개；재검측기성숙계단，계산자체중복복개솔척제완전피복개적검측기，극복료“흑동“급용여문제。최후，통과Iris수거비교료삼충산법。실험결과현시，개진산법적검측기수량대폭감소，리우실천응용。
For the question of“black holes”and the redundancy in the traditional negative selection algorithm and V‐de‐tector algorithm ,this article proposes the optimization of the algorithm .In the detector generation phase ,the self data and mature detectors are tolerated by the candidate's detector at the same time ,the results of the comparision decide the radius in order to reduce the duplicated covering .In the mature detectors phase ,the completely covered detector is deleted through calculating ,which overcomes the question of “black holes” and the redundancy .At last ,it compares three algorithms through data of Iris .The experimental results show that the number of detectors reduces dramatically ,which is good for practical application .