路基工程Issue(3)：17-19,3.
基于改进神经网络的路基沉降预测
Subgrade Settlement Forecast Based on Improved Neutral Network
摘要
Abstract
In order to overcome the defects in subgrade settlement forecast method and shortcomings in traditional BP neutral network, Levenberg-Marquardt arithmetic is used to improve convergence of BP neutral network so as to establish the model for subgrade settlement forecast. Combined with the measured data of subgrade settlement on Tianjin-Qinhuangdao PDL, this optimized model and traditional forecast model is compared. The result shows that the improved BP neutral network is of highest accuracy and good performance in subgrade settlement forecast.关键词
路基/预测模型/沉降预测/神经网络Key words
subgrade/forecast model/settlement forecast/neutral network分类
建筑与水利引用本文复制引用
蒲兴波,魏静,钱耀峰,李军昌..基于改进神经网络的路基沉降预测[J].路基工程,2012,(3):17-19,3.基金项目
中央高校基本科研业务费专项资金项目 （）
轨道工程北京市重点实验室 （）
2011年大学生创新实验项目 （）
