中文摘要：

为进一步提升网络入侵检测效果，提出一种融合FAST特征选择与自适应二进制量子引力搜索支持向量机的（FAST-ABQGSA-SVM）网络入侵检测算法。利用FAST算法过滤掉原始特征集中冗余无关的特征形成候选特征子集，基于组合优化策略采用自适应二进制量子引力搜索算法对候选特征子集与SVM分类器参数进行组合优化。在ABQGSA反复学习寻优过程中，采取动态自适应波动式调整策略更新量子旋转角以平衡算法全局搜索能力和局部搜索能力；同时为提升算法的自适应变异能力，设计与进化程度及个体适应度值相关的自适应变异概率，当种群进化出现停滞时及时引入量子位离散交叉操作帮助种群摆脱局部极值。通过KDD CUP 99仿真实验表明，所提出的FAST-ABQGSA-SVM算法较其他同类型检测算法具有更好的鲁棒性、学习精度以及检测效果。

英文摘要：

Aiming at improving the effect of network intrusion detection, this paper introduced an intrusion detection algorithm based on FAST feature selection algorithm, adaptive binary quantum-inspired gravitational search algorithm and support vector machine（FAST-ABQGSA-SVM）.It used FAST algorithm to remove some irrelevant features to form the candidate feature subset.It used BQGSA to solve the problem of combinatorial optimization to obtain the optimal feature subset and the classifier parameters.During the process of repeated learning optimization, using a dynamic adaptive adjustment strategy to update the quantum rotation angle, it could balance the global search ability and local search ability.Designing the adaptive mutation pro-bability based on population evolutionary degree and individual fitness value, it improved the adaptive mutation ability.Using Q-bit crossover operation when the population evolutionary stagnation occurred, it could help the population get rid of local extremum.Simulation results by KDD CUP 99 show that the proposed FAST-ABQGSA-SVM algorithm has better robustness, learning accuracy and detection performance compared with other similar detection algorithms.