论文部分内容阅读
介绍有别于传统意义上的反演分析方法,利用训练好的神经网络能够替代有限元进行计算,将遗传算法和神经网络结合,既能利用神经网络的非线性映射、网络推理和预测功能,又利用了遗传算法的全局优化特性,有效地克服了有限元计算反分析效率低及常规的优化算法与初值关系密切的弊端,并结合武汉阳逻长江公路大桥南锚超深基坑工程,较为详细地介绍其实施过程。
The introduction is different from the traditional method of inversion analysis. The trained neural network can be used to replace the finite element calculation. Combining the genetic algorithm with the neural network, the neural network can not only utilize the nonlinear mapping, network reasoning and prediction functions, But also take advantage of the global optimization characteristics of genetic algorithm to effectively overcome the drawbacks of low efficiency and conventional optimization algorithms of finite element calculation and initial value. Combining with the South anchored ultra deep foundation pit project of Yangluo Yangtze River Bridge in Wuhan, A more detailed description of its implementation process.