CIESC Journal ›› 2020, Vol. 71 ›› Issue (3): 1080-1087.DOI: 10.11949/0438-1157.20191495

• Process system engineering • Previous Articles     Next Articles

Intrusion detection of industrial control system based on grey wolf optimization integrated random black hole

Zhiqiang GENG1,2(),Rongfu ZENG1,2,Yuan XU1,2,Yongming HAN1,2(),Xiangbai GU1,2,3   

  1. 1.College of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, China
    2.Engineering Research Center of Intelligent PSE, Ministry of Education of China, Beijing 100029, China
    3.Sinopec Engineering Group Co. , Ltd. , Beijing 100029, China
  • Received:2019-12-09 Revised:2019-12-13 Online:2020-03-05 Published:2020-03-05
  • Contact: Yongming HAN

融合灰狼优化算法在工控系统入侵检测中的应用

耿志强1,2(),曾荣甫1,2,徐圆1,2,韩永明1,2(),顾祥柏1,2,3   

  1. 1.北京化工大学信息科学与技术学院,北京 100029
    2.智能过程系统工程教育部工程研究中心,北京 100029
    3.中石化炼化工程(集团)股份有限公司,北京 100029
  • 通讯作者: 韩永明
  • 作者简介:耿志强gengzhiqiang@mail.buct.edu.cn
  • 基金资助:
    国家重点研发计划项目(2018YFB0803501);国家自然科学基金项目(61673046);中央高校基本科研业务费专项资金(XK1802-4)

Abstract:

Aiming at the characteristics of large data volume and high dimensions in the current industrial control system, a grey wolf optimization integrated random black hole (RBHGWO) algorithm incorporating a random black hole (RBH) strategy is proposed. When the wolf group updates the position of the next generation grey wolf, the proposed algorithm simulates the attraction of black holes, so that the individual in the wolf group can move faster towards the current global optimal solution, and enhances the convergence speed of the proposed algorithm. Meanwhile, individuals are randomly attracted by black holes, which maintain the local search ability of the proposed algorithm. Compared with particle swarm optimization (PSO), random black hole particle swarm optimization (RBHPSO), GWO algorithm and survival of fitness grey wolf optimization (SFGWO) algorithm using test functions, the experimental results show that the RBHGWO algorithm has fast convergence speed and excellent convergence accuracy. Moreover, based on the data set of Tennessee-Eastman (TE) simulation platform, the situation of industrial control systems is simulated by attacking from the covert intrusion. The experimental results show that the RBHGWO algorithm has obvious advantages in convergence accuracy, iteration speed and stability in the feature selection of intrusion detection of industrial control systems.

Key words: industrial control system, intrusion detection, feature selection, algorithm, optimization, simulation

摘要:

针对当前工控系统中数据体量大、维度高的特点,提出了一种融合随机黑洞(random black hole, RBH)策略的灰狼优化(grey wolf optimization integrated random black hole, RBHGWO)算法。该方法在更新下一代灰狼位置时,模拟黑洞的吸引方式,让狼群中的个体能够更快地向着当前全局最优解移动,增强了算法的收敛速度;同时个体以随机策略被黑洞吸引,保持了算法的局部搜索能力。通过优化算法测试函数验证,RBHGWO算法与粒子群优化(particle swarm optimization, PSO)算法、随机黑洞粒子群优化(particle swarm optimization integrated random black hole, RBHPSO)算法、GWO算法和优胜劣汰的灰狼优化(survival of fitness grey wolf optimization, SFGWO)算法进行了实验对比。结果表明,RBHGWO算法具有较快的收敛速度和较好的寻优精度。同时以田纳西-伊斯曼(Tennessee-Eastman, TE)数据集为基础进行仿真实验,结果表明该算法应用于在工控系统入侵检测的特征选择中,其收敛精度、迭代速度以及稳定性都有明显优势。

关键词: 工业控制系统, 入侵检测, 特征选择, 算法, 优化, 模拟

CLC Number: