工业控制网络中入侵波动抑制模型仿真  

Simulation of Intrusion Fluctuation Suppression Model in Industrial Control Network

作　　者：杨斌[1] 段娜[1] 赵军民[1] 

机构地区：[1]河南城建学院计算机科学与工程学院,河南平顶山467036

出　　处：《计算机仿真》2015年第9期295-298,共4页Computer Simulation

基　　金：国家语委"十二五"科研规划重点科研项目(ZDI125-23)

摘　　要：针对工业控制网络中入侵行为产生的波动,传统的抑制方法在工作过程中,由于入侵波动具有隐匿性、无序性与不稳定性,加之网络噪声的干扰,造成抑制源过多,无法有效辨认,存在效果差、抑制效率低的缺陷。提出利用贝叶斯推理与RBF神经网络优化算法相融合的工业控制网络中入侵波动的抑制方法。采用贝叶斯相关理论建立网络入侵检测的数学模型,实现网络入侵的检测,对网络入侵的波动范围加以确定,依据RBF神经网络优化算法分别计算网络的误差信号及误差阀值,并进行比较,以此作为算法终止的判定条件,通过对时间进行周期性设置从而实现连接权值的实时修正,输出最佳入侵波动的抑制结果。实验结果表明,采用改进算法进行工业控制网络中入侵波动的抑制,能够达到满意的抑制效果,有效的提高网络入侵检测的准确率与检测效率,提高了工业控制网络的安全性与稳定性。The paperis proposed an inhibition method for intrusion fluctuation in the industrial control network based on the integration of Bayesian inference and RBF neural network optimization algorithm. Theory of Bayesian was used to establish the mathematical model of network intrusion detection, realize the network intrusion detection and determine the network intrusion range. On the basis of RBF neural network optimization algorithm, the network error signal and the error threshold were calculated and compared, which were taken as judging conditions for the ter- mination of the algorithm. Through the periodic set on time, wet realized the real-time correction of the connection weights, and output the optimal inhibition results of intrusion fluctuations. Experimental results show that using the improved algorithm for intrusion fluctuation inhibition of industrial control network can achieve satisfactory effect, ef- fectively improve the accuracy and detection efficiency of network intrusion detection, and then enhance the security and stability of industrial control network.

关 键 词：工业控制网络 入侵波动抑制 贝叶斯推理 神经网络 

分 类 号：TH128[机械工程—机械设计及理论]