基于遗传神经网络的再生沥青混合料性能预测研究

以matlab为平台,分别应用BP神经网络和遗传算法优化的BP神经网络对再生沥青混合料的性能进行预测。以旧料掺量、油石比等8个影响因素作为输入层,以动稳定度、残留稳定度等5个性能指标作为输出层,将28组归一化处理后的试验数据进行神经网络的训练、验证和测试。结果表明:遗传算法优化的BP神经网络预测表现出更加精准的预测效果。将遗传算法优化的BP神经网络应用于工程实践中,再生沥青混合料性能预测可以大大提高试验科学性和预见性。

Prediction of Performance of Recycled Asphalt Mixture Based on Genetic Neural Network

Matlab is used as a platform to predict the performance of recycled asphalt mixture using BP neural network and genetic algorithm optimized BP neural network.Eight input factors such as old material content and oil-cement ratio were used as the input layer,and five performance indicators such as dynamic stability and residual stability were taken as output layers.The 28 sets of normalized test data were used to train the neural network.,verification and testing.The results show that the BP neural network optimized by genetic algorithm shows a more accurate prediction effect.The application of the BP neural network optimized by genetic algorithm in the engineering practice to predict the performance of recycled asphalt mixture can greatly improve the scientific and predictability of the experiment.