【摘要】 为了提高计算机网络流量在线预测的精度,根据网络流量的时变、非线性特点,利用人工神经网络在非线性建模方面的优势建立了一个时间相关的神经网络预测模型.给出了相应的模拟在线神经网络预测模型结构,详细分析了改进的PSO算法在神经网络权值学习中的应用,并对实际的网络流量数据进行了仿真分析.Matlab仿真结果表明:该模型具有较好的预测效果,相对于BP神经网络模型具有更快的收敛速度和更好的自适应性.更多还原

【Abstract】 In order to improve the precision of computer network traffic prediction,and according to the traffic’s time-varying and nonlinear characteristics,a time-related forecasting model with artificial neural network(ANN) is established,for ANN has huge advantages in nonlinear modeling.The prediction model structure of time-related ANN is presented at first,then the improved particle swarm optimization(PSO) algorithm in neural network weight learning is introduced in detail,and the prediction model is applied to the analysis of the actual network traffic data.The simulation results of Matlab indicate that the model has faster convergence speed and better adaptability than BP neural network model,at the same time,the forecast effect is good.更多还原