GA-BP神经网络在偏压隧道拱顶沉降预测的应用

Application of GA-BP neural network in vault settlement prediction of the bias tunnel

运用组合优化的思想,基于遗传算法(GA)优化BP神经网络,将优化后的神经网络(GA-BP)用来预测偏压隧道拱顶沉降量。通过对比分析BP神经网络和经过遗传算法优化过的神经网络的预测结果数据表明,在偏压隧道拱顶沉降预测中,后者比前者性能更好,更能精确的预测偏压隧道拱顶沉降量。

By using the combination optimization,based on the BP neural network of genetic algorithm,the optimized neural network（ GA-BP） was used to predict the vault settlement of the bias tunnel. Through the comparative analysis about the prediction results of general BP neural network and optimized neural network by genetic algorithm,the experimental results show that,in the vault settlement prediction of bias tunnel,the latter performs better than the former and can predict for the vault settlement of the bias tunnel more accurately.