一种改进QPSO优化BP网络的入侵检测算法

为了较好克服量子粒子群算法存在早熟收敛的缺点，在分析算法参数和流程的基础上，提出了一种带变异操作的改进量子粒子群优化算法。针对传统BP算法易于陷入局部极小的不足，将改进的算法应用到BP神经网络的学习过程中，修正BP网络的权值和阈值，提高其收敛性能。并将优化的BP神经网络模型应用于入侵检测中，用标准入侵检测数据对基于不同算法的BP网络进行仿真实验比较。实验结果表明，改进后的BP算法迭代次数少，收敛速度有所提高，在一定程度上提高了入侵检测率。

An Intrusion Detection Algorithm of BP Network Optimized by Improved QPSO

In order to overcome the shortcomings of the quantum particle swarm optimization algorithm better, which is precocious conver- gence, on the basis of analyzing algorithm parameters and processes, an improved quantum particle swarm optimization algorithm with mutation operation has bnea proposed. Because the traditional BP algorithm is easy to fall into local minima,the improved algorithm is ap- plied in the learning process of the BP neural network to correct weights and threshold of BP network, and improve the convergence per- formance. The optimized BP neural network is used in the intrusion detection, and simulation experiments on BP network with different algorithm is made with standard intrusion detection data. The results show that the improved BP algorithm has less number of iterations, the convergence rate has increased,improving the intrusion detection rate too.