摘要
对Castro和T imm is提出的带变异的阴性选择算法进行适当改进,用于故障模式的离线训练。对Cas-tro和Zuben提出的aiN et免疫网络算法进行适当改进,并用于故障的在线检测。实验表明,由于采用了聚类的思想,构造的自我集和非我集能较全面反映系统在正常与不正常两种模式下的全貌,提高了在线检测准确率。
In this paper, negative selection algorithm with mutation proposed by Castro and Timmis is improved and used in the off-line training of fault modes. The aiNet immune network algorithm proposed by Castro and Zuben is improved and used in online detection. Diagnosis case shows that the normal pattern and theabnormal pattern can be covered with the ＇self＇ set and the ＇non-self＇ set largely because of clustering. The accuracy of detection is improved.
出处
《电脑开发与应用》
2007年第3期32-34,共3页
Computer Development & Applications
关键词
故障检测
阴性选择算法
免疫网络
fault detection, negative selection algorithm, immune network
