摘要
建立了以改进Powell法优化的前馈型神经网络模型 ,模型既具有强大的函数逼近功能 ,又克服了传统神经网络优化方法的缺点。将神经网络模型用于承压水漏斗的动态水位预报 ,实例表明 :模型预报效果较好。
A feedforward neural network model optimized by the modified Powell method is established. The model is of strong function of approximation and capable of overcoming the disadvantages of conventional feedforward neural network optimization method. So, it is used to forecast the confined water funnel level. The case shows that the forecasting results of the model preferably.
出处
《江苏地质》
2002年第1期19-21,共3页
Jiangsu Geology
