基于遗传算法和遗传神经网络算法的堆石料参数反演分析研究

面板堆石坝应力变形计算过程相当费时费力,如果直接调用有限元计算程序进行反演计算,难度相当大,效率也非常低.利用遗传算法优化BP神经网络权值与阈值,建立遗传神经网络模型代替堆石坝的有限元计算程序以提高反演计算效率,同时利用遗传优化算法全局搜索功能寻找使遗传神经网络模拟值和实测值之间误差最小的最优参数组,并通过MATLAB实现基于遗传算法和遗传神经网络算法的堆石料参数反演分析,反演结果表明该算法能够很好地提高反分析效率及准确性.

A Back-analysis of Rockfills Based on Genetic Algorithm and Genetic Algorithm Neural Networks

In the finite element calculation of face rockfill dams,it is often necessary to do complicated,time-consuming and energy-consumering design calculations.This paper uses genetic algorithm optimization BP neural network weights and thresholds and establishes the genetic neural network model for finite element simulation of rockfill dams,then combining the principle of genetic algorithm optimization algorithm.Based on genetic algorithm optimization and genetic algorithm neural network,A back-analysis of rockf...