摘要
为准确预测滑坡变形趋势,针对黄土滑坡位移变化复杂的非线性特征,提出基于遗传算法与粒子群算法优化的Elman神经网络模型(GA-PSO-Elman)。顾及Elman神经网络结构参数的随机性可能陷入局部最优解,将遗传算法(GA)较强的全局搜索能力与粒子群算法(PSO)的局部搜索能力相结合,优化预测模型结构参数的权值阈值,提高预测预报精度与收敛速度。将该模型应用于甘肃临夏黑方台滑坡变形预测,结果表明,相较于传统BP神经网络和单一Elman神经网络,新模型具有更高的精度和稳定性。进一步考虑影响滑坡的诸多因素,将湿度、降水等影响因子加入各个训练模型,GA-PSO-Elman模型的学习速度和收敛速度进一步提高,从而有效提高变形预测精度。
In order to accurately predict the landslides deformation trend,aiming at the complex nonlinear characteristics of loess landslide displacement change,we propose an Elman neural network model based on genetic algorithm and particle swarm optimization algorithm(GA-PSO-Elman).Considering the problem that Elman neural network may fall into local optimal solution due to the randomness of structural parameters,we combine the strong global search ability of genetic algorithm(GA)with the local search ability of particle swarm optimization(PSO)to optimize the weight and threshold of structural parameters of the prediction model,and improve the prediction accuracy and convergence speed.The model is applied to the deformation prediction of Heifangtai landslide in Linxia,Gansu Province.The results show that the new model has better accuracy and stability than traditional BP neural network and single Elman neural network.Considering many factors affecting landslide,and adding humidity and precipitation to each training model,we improve the learning speed and convergence speed of GA-PSO-Elman model,thus effectively improving the accuracy of deformation prediction.
作者
王志彪
赵丽华
WANG Zhibiao;ZHAO Lihua(School of Geological Engineering and Geomatics,Chang’an University,Xi’an 710054,China)
出处
《大地测量与地球动力学》
CSCD
北大核心
2023年第7期679-684,共6页
Journal of Geodesy and Geodynamics
基金
国家重大科研仪器研制项目(42127802)。
关键词
ELMAN神经网络
遗传算法
粒子群算法
预测
Elman neural network
genetic algorithm
particle swarm optimization
prediction
作者简介
第一作者:王志彪,硕士生,主要从事滑坡预测预报算法研究,E-mail:wangzb_97@163.com。;通讯作者:赵丽华,博士,副教授,主要从事大地测量数据处理方法教学与研究,E-mail:zhaolih@chd.edu.cn。
