摘要
为解决智能电网运行入侵检测效率慢及入侵检测精度较低等问题,提出基于量子甲虫群算法优化的极限学习机模型。通过构建量子甲虫群优化算法,并引入量子力学,结合甲虫触角搜索和粒子群优化的优点,以进一步提高极限学习机算法入侵收敛性能,降低极限学习机的计算复杂度和训练时间。结果表明:随着迭代次数的增加,入侵检测测试误差逐渐减小,最小误差率为1.1%。所提出的极限学习机算法的准确率、平均F值和攻击准确率分别为95.82%、95.90%和95.16%。与随机森林算法相比,极限学习机可以有效提高智能电网运行入侵检测的准确性、检测率、攻击准确率,降低误报率,算法可满足实际智能电网运行入侵检测。
To further solve the problems of slow intrusion detection efficiency and low intrusion detection accuracy in smart grid operation,an extreme learning machine model based on quantum beetle swarm optimization algorithm is proposed.By constructing a quantum beetle swarm optimization algorithm and introducing quantum mechanics,combining the advantages of beetle antenna search and particle swarm optimization,the intrusion convergence performance of extreme learning machine algorithm is further improved,and the computational complexity and training time of extreme learning machine are reduced.The experimental results show that as the number of iterations increases,the error of intrusion detection testing gradually decreases,with a minimum error rate of 1.1%.The accuracy,average F value and attack accuracy of the proposed extreme learning machine algorithm are 95.82%,95.90%and 95.16%respectively.Compared with random forest algorithm,extreme learning machine can effectively improve the accuracy,detection rate and attack accuracy of intrusion detection in smart grid operation,and reduce false alarm rate,and extreme learning machine algorithm can meet the actual intrusion detection in smart grid operation.
作者
梁林森
LIANG Linsen(Information Centre,Guangzhou Power Supply Bureau,Guangdong Power Grid Co.,Ltd.,Guangzhou 510000,China)
出处
《粘接》
CAS
2023年第8期185-188,共4页
Adhesion
关键词
极限学习机
智能电网
入侵
检测研究
extreme learning machine
smart grid
invasion
detection research
作者简介
梁林森(1982-),男,硕士,高级工程师,研究方向:信息技术等,E-mail:llintao3482@126.com。
