针对人工免疫系统中V-detector否定选择算法造成的检测器集合黑洞和检测器高重叠率等问题,借鉴生物免疫系统对免疫细胞的调节机制,提出了V-detector优化算法。该算法从父代产生候选检测器子代并通过检测器之间以及检测器与自体集合之间的亲和力对比更新检测器集合,使得检测器集合对非自体空间的覆盖更加合理。通过二维仿真实验和KDDCUP99数据集实验测试,经优化后的检测器集合对非自体空间的覆盖性能有了显著提高,有效提高了系统的检测性能。
An optimization algorithm for variable-coverage detectors (V-detectors) was designed based on the immunecell regulation mechanism in biology immune systems to solve the problems of V-detector hole and high V-detector overlapping of the V-detector algorithm, a real-valued negative selection algorithm with V-detectors. The algorithm updates the detector set by the candidates generated from their parents and the affinity comparison to improve detectors' distribution performance. It was tested by the synthetic data and the KDD CUP 99 data set. The results show that the optimized detectors can increase the efficiency of detectors' distribution and improve the system' s detection performance.