摘要
针对基于网络误用入侵检测模型的入侵特征库存在构建困难、自适应差的缺点,提出了一种基于量子遗传算法的入侵特征库优化算法。首先通过提取网络协议中容易被攻击和修改的特征值,经组合和编码后构成算法的初始种群。然后以检测率和误警率为评价指标设计适应度函数,利用量子旋转门更新染色体,随着算法的运行逐代优化种群。实验仿真结果表明:该算法在寻优能力与收敛速度上均优于对应的遗传算法;经该算法优化后的种群,检测能力强、自适应性好。
Aimed at the shortcomings that there are difficulties of constructing and bad self adaptability in the intrusion signature sets based on the network misuse intrusion detection model,the method of optimizing intrusion signature sets with quantum genetic algorithm is proposed.Firstly,the features which are easily attacked and modified are extracted from network protocol,and the initial population is produced by combining and coding.Then,the fitness function based on the evaluation indexes of the detection rate and false-alarm-rate is designed,the chromosomes are updated with the quantum rotation gate,and the population is optimized generation by generation as the algorithm runs.Simulation results show that the QGA is superior to the corresponding GA in search ability and convergence rate,and the population optimized by QGA had better detection ability and self adaptability.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第12期2933-2935,F0003,共4页
Computer Engineering and Design
基金
浙江省教育厅科研基金项目(Y200909706)
关键词
入侵检测
量子遗传算法
入侵特征
检测率
误警率
intrusion detection
quantum genetic algorithm
intrusion signature
detection rate
false-alarm-rate