长春师范大学学报（自然科学版）
長春師範大學學報（自然科學版）
장춘사범대학학보（자연과학판）
Journal of Changchun Normal University (Natural Science)
2014年
4期
47-50
,共4页
网络入侵检测%神经网络%禁忌算法%优化
網絡入侵檢測%神經網絡%禁忌算法%優化
망락입침검측%신경망락%금기산법%우화
network intrusion detection%neural network%tabu algorithm%optimization
随着互联网的发展和普及，传统网络入侵防范方法如防火墙、数据加密等已经很难保证系统和网络资源的安全。为此，本文设计了基于改进禁忌算法和神经网络的网络入侵检测方法。首先建立三层的 BP 神经网络模型用于实现入侵检测。然后通过 BP 反向传播算法获取网络的权值和阀值等参数，并设计了一种基于双禁忌表的改进禁忌优化算法，采用此改进的禁忌优化算法对 BP算法优化得到的权值和阀值进行进一步寻优。最后，将禁忌算法优化后的神经网络用于网络入侵检测。仿真实验表明，此方法能够有效地实现网络入侵检测，具有较快的收敛速度和较高的检测率，是一种适合网络入侵检测的可行方法。
隨著互聯網的髮展和普及，傳統網絡入侵防範方法如防火牆、數據加密等已經很難保證繫統和網絡資源的安全。為此，本文設計瞭基于改進禁忌算法和神經網絡的網絡入侵檢測方法。首先建立三層的 BP 神經網絡模型用于實現入侵檢測。然後通過 BP 反嚮傳播算法穫取網絡的權值和閥值等參數，併設計瞭一種基于雙禁忌錶的改進禁忌優化算法，採用此改進的禁忌優化算法對 BP算法優化得到的權值和閥值進行進一步尋優。最後，將禁忌算法優化後的神經網絡用于網絡入侵檢測。倣真實驗錶明，此方法能夠有效地實現網絡入侵檢測，具有較快的收斂速度和較高的檢測率，是一種適閤網絡入侵檢測的可行方法。
수착호련망적발전화보급，전통망락입침방범방법여방화장、수거가밀등이경흔난보증계통화망락자원적안전。위차，본문설계료기우개진금기산법화신경망락적망락입침검측방법。수선건립삼층적 BP 신경망락모형용우실현입침검측。연후통과 BP 반향전파산법획취망락적권치화벌치등삼수，병설계료일충기우쌍금기표적개진금기우화산법，채용차개진적금기우화산법대 BP산법우화득도적권치화벌치진행진일보심우。최후，장금기산법우화후적신경망락용우망락입침검측。방진실험표명，차방법능구유효지실현망락입침검측，구유교쾌적수렴속도화교고적검측솔，시일충괄합망락입침검측적가행방법。
With the further development of the Internet and the popularization of network, the traditional network detection methods such as firewall and data encryption cannot guarantee the security of system and network resource. Therefore, the network intrusion method based on the improved tabu algorithm and neural network is proposed in this paper. Firstly, the three - layer BP model for network intru-sion detection is set up, then the BP back propagation algorithm is used to obtain the parameters such as weight and threshold, and an improved tabu algorithm based on double tabu table is designed and used to optimize the parameters such as weight and threshold. Final-ly, the tabu algorism optimizing neural network is put forward to detect network intrusion. The simulation experiments show that this method can effectively realize the network intrusion detection. The method has faster convergence speed and higher detection rate, so it is suitable for a feasible way to detect network intrusion.
