土工基础

土工基础 ›› 2022, Vol. 36 ›› Issue (6): 907-910.

• 专题论述 • 上一篇    下一篇

基于遗传算法优化BP神经网络的压缩系数预测研究

欧阳磊1，邓希萍2，何利军1，刘家印1，陈 鹏1   

  1. (1.南昌航空大学 土木建筑学院，南昌 330063；2.萍乡市规划勘察设计院，江西萍乡 337000)
  • 收稿日期:2021-05-30 修回日期:2021-06-11 出版日期:2022-12-31 发布日期:2022-12-28
  • 通讯作者: 何利军（1977-），男，博士，讲师，主要研究方向为岩土工程。
  • 作者简介:欧阳磊(1996-)，男，在读硕士研究生，主要研究方向为工程数值模拟与计算。
  • 基金资助:
    广西防灾减灾与工程安全重点实验室开放课题(2016ZDK013)；广西大学工程防灾与结构安全教育部重点实验室开放课题(2016ZDK013)

Prediction of Soil Compression Coefficient by Genetic Algorithm Optimized BP Neural Network

OUYANG Lei1, DENG Xiping2, HE Lijun1, LIU Jiayin1， CHEN Peng1   

  1. (1.School of Civil Engineering, Nanchang Hangkong University, Nanchang 330063；
    2.Pingxiang City Planning Survey and Design Institute, Pingxiang 337000)
  • Received:2021-05-30 Revised:2021-06-11 Online:2022-12-31 Published:2022-12-28

摘要： 压缩系数越大，土的压缩性就会越高，土的压缩性过高容易发生地质灾害而造成不可扭转的损失，预测压缩系数对于预防地质灾害意义深远。基于BP神经网络原理，以常规物理参数为输入向量，以压缩系数为输出向量，利用BP神经网络具有非线性映射能力强的特点以及遗传算法可以很好地解决BP神经网络易陷入局部最优值问题的优势，在构建基于BP神经网络的压缩系数预测模型的基础上利用遗传算法优化该模型，优化后预测性能提高很多，最大相对误差和最小相对误差分别为22.38%和-1.13%，对于岩土工程来说，误差在25% 之内是可接受范围，故预测模型在一定程度上可以作为替代工具预测压缩系数。

关键词: 压缩系数, BP神经网络, 遗传算法, 预测

Abstract:

The larger the soil compression coefficient, the higher the compressibility of the soil. Excessive compressibility of the soil is likely to cause geological hazards and cause irreversible losses. Predicting the soil compression coefficient has far-reaching significance for the prevention and minimization of geological hazards. Based on the principles of the BP neural network, with conventional physical parameters as input vector and compression coefficient as output vector, the use of the BP neural network with a strong nonlinear mapping ability and genetic algorithm can solve the problem of the BP neural network easily falling into the local optimal value. Based on constructing a compression coefficient prediction model by using the BP neural network, the genetic algorithm is used to optimize the model. After the optimization, the prediction performance is greatly improved. The maximum relative error and the minimum relative error are 22.38% and 1.13%, respectively. In engineering terms, an error within 25% is an acceptable range, so the prediction model can be used as an alternative tool to predict the compression factor to a certain extent.

Key words: Soil Compression Coefficient, BP Neural Network, Genetic Algorithm, Prediction

中图分类号: