基于改进的灰狼算法优化BP神经网络的人侵检测方法

打开文本图片集

Intrusion detection method based on improved grey wolf optimization algorithm optimized BPneural network

PENG Qingyuan1， WANG Xiaofeng，²，TANG Ao1，HUA Yingying1，HE Fei'，LIU Jianping1，² （1.SchoolofComputerScienceandEngineering，NorthMinzu University，Yinchuan75oo21，China; 2.KeyLabelsassi

Abstract：Network securityissuesare becoming moreand more prominent in today'sworld.Theintrusiondetection technologyhasbeenrapidlydevelopedasanimportantpartinthefieldofnetworksecurity.Atpresent，BPneuralnetworkis widelyusedinintrusiondetection.However，thweightseletingofthetaditioalBnuraletworkisiaccurate，tsleaing eficiencyislowanditispronetofalingintolocalminima.Fortheaboveshortcomings，anintrusiondetectionmethodbasedon theimproved greywolfoptimization（IGWO）algorithmoptimizedBPneuralnetwork isproposed.TheIGWOalgorithmextends thesearcrangeof thewolf pack bychanging the linearcontrolparametersandadingtheinversecotangentinertia weight strategyinthegraywolfpositionupdateformulatoavoidfalingintothelocaloptimalsolution.Theimprovedalgorithmisused tooptimizetheinitialweightsandthresholdvaluesoftheBPneuralnetwork，andtheoptimizedBPneuralnetworkisappliedto intrusiondetection.TheexperimentalresultsshowthattheIGWOalgorithmhasbeterstability，optimizing eficiencyand optimizingaccuracy，andtheimprovedintrusiondetectionmethodisnotprone tofaling intolocalminima，hasstrong generalization ability，and has high prediction accuracy and reliability.

Keywords：nonlinear control parameter;inertia weight;GWOalgorithm;BPneural network；intrusiondetection;network security

0 引言

随着网络技术的发展，计算机网络的安全性受到越来越多的关注。（剩余10024字）