中文摘要：

从力学作用机理出发探讨堆载预压效应下路基沉降与堆载的关系,以径向基函数神经网络为建模工具,以现场监测的沉降资料及堆载记录为基础,构建堆载沉降的神经网络监测模型的输入层影响因子和包含多个测点沉降的输出层.同时,针对堆载沉降特点,对网络的计算中心采用专门预选方案,进而以模糊C均值聚类算法确定最终计算中心,并以实测资料建立该监测模型实例.结果表明,所得模型对堆载作用的揭示以及多测点的沉降预测取得了满意效果.

英文摘要：

To analyze and forecast road foundation settlement caused by heaping pre-compaction,the relationship between settlement and heaping process were studied according to the heredity-creep theory.Radial basis function（RBF） artificial neural network was used to establish a monitoring model.In the model,the input units aim at heaping settlement and the output units including multi-point settlement are formed based on surveyed settlement in-site and heaping recordation.Moreover,a pre-select RBF center method was presented considering heaping settlement character.It works with fuzzy C-means algorithm（FCM） together to decide final centers.Model instance established in above way with monitoring data describes the heaping effect well,and with good training and forecast results.