基于遗传算法优化的BP神经网络预测混凝土抗压强度

Research on the Prediction of Concrete Compressive Strength Based on BP Neural Network Optimized by Genetic Algorithm

针对传统BP神经网络预测能力的不足,采用遗传算法优化的BP神经网络预测混凝土28d抗压强度,并建立了GA-ANN预测模型。一方面对模型输入项进行灰色关联分析,找出对强度影响的重点关联因素;另一方面,研究了模型在不同输入项组合、不同训练组和输出组数、不同网络隐含层节点数的操作条件下,其预测精度有何变化,找出了最优的上述操作条件,减少了预测误差。

For the limited predictive ability of BP neural network medel, article uses the genetic algorithm （GA） to optimize BP neural network model and establishes the GA - ANN prediction model to predict the 28 - days concrete compressive strength. On the one hand, article uses gray relative analysis method to select the key factors of the concrete compressive strength; on the other hand, the influences of the operating conditions, like the different input group, different training and output group count, different nodes of hidden layer are also discussed. To reduce the error of the prediction, the best operating conditions are confirmed.