基于神经网络理论的沥青路面结构四层体系弯沉的拟合研究

A Study on Fitting of Deflection of Four layer Elastic-layered System for Asphalt Pavement Based on the Theory of Neural Networks

根据弹性层状体系理论 ,采用三层双并联前向神经网络 ( DPFNN) ,研究了沥青路面结构四层体系理论弯沉的拟合计算问题 ,通过大量的计算与验证 ,训练好的网络精度高、速度快、泛化能力强 ,且易于实现 ,可应用于工程设计与验算。

Based on the theory of elastic layered system,the fitting problem is researched for calculating the deflection of four layer elastic layered system in asphalt pavement by using the three layer double parallel feedforward neural network (DPFNN). Through a great deal of computations and checks,the trained network is high in accuracy,quick in speed,good in expansion and easy to achieve.The network can be used in the engineering designs and checks.