基于BP-PSO联合算法的沥青混合料空隙率反演计算

Inversion calculation of void ratio of asphalt mixture based on BP-PSO joint algorithm

为提高沥青混合料空隙率反演计算效率,提出一种基于BP神经网络和PSO联合算法的沥青混合料空隙率反演计算方法.通过建立探地雷达电磁波在沥青路面结构层中的传播仿真模型,利用BP神经网络对仿真样本进行训练,得到能够反映输入和输出关系的回归曲面.利用PSO算法对回归曲面进行寻优,得到最优解.将该联合算法应用于沥青混合料空隙率反演计算中,反演计算结果表明:该联合算法能够在满足计算精度的条件下大幅提高计算效率,证明应用该联合算法反演计算沥青混合料空隙率的可行性和有效性.

In order to improve the inversion calculation efficiency of the asphalt mixture void ratio, an inversion calculation of asphalt mixture void ratio based on BP neural network and PSO joint algorithm is proposed. By establishing the propagation simulation model of ground penetrating radar electromagnetic wave in the structural layer of asphalt pavement, the BP neural network is used to train the simulation samples, and the regression surface that can reflect the relationship between input and output is obtained. PSO algorithm is used to calculate and obtain the optimal solution of the regression surface;the joint algorithm is applied to the inversion calculation of asphalt mixture void ratio, and the results show that: This joint algorithm can greatly improve the efficiency of calculation under the condition of guaranteeing the accuracy of calculation, proving the feasibility and effectiveness of the joint algorithm on inversion calculation efficiency of the asphalt mixture void ratio.