###
DOI:
有色金属工程:2022,(1):-
←前一篇   |   后一篇→
本文二维码信息
码上扫一扫!
基于IPSO-ELM的边坡稳定性分析
赵允坤, 胡 军, 杨 斌
(辽宁科技大学土木工程学院)
Slope Stability Analysis Based on IPSO-ELM
ZHAO Yunkun, HU Jun, YANG Bin
(School of Civil Engineering University of Science and Technology LiaoNing,Anshan)
摘要
图/表
参考文献
相似文献
本文已被:浏览 979次   下载 779
投稿时间:2021-07-01    修订日期:2021-07-08
中文摘要: 边坡稳定性受多种复杂因素影响,传统算法很难得到高精度预测结果,为了及时准确地对边坡稳定性做出可靠性分析,提出了改进粒子群优化极限学习机(IPSO-ELM)模型并应用于边坡稳定性预测实例中。首先在粒子群算法(PSO)的基础上,为克服在寻优过程中易出现局部最优的问题,引入自适应权重法,将改进粒子群算法(IPSO)对极限学习机(ELM)的输入权值和隐层偏置进行优化,大大提高了ELM模型的泛化能力和预测精度,然后将IPSO-ELM模型、PSO-ELM模型和ELM模型的预测值与真实值对比分析,结果表明IPSO-ELM模型预测值趋近于真实值,有较高的预测精度,验证了IPSO-ELM模型在评价边坡稳定性中的可行性和有效性。
Abstract:Slope stability is affected by a variety of complex factors, and it is difficult for traditional algorithms to obtain high precision prediction results. In order to timely and accurately analyze the reliability of slope stability, an improved particle swarm optimization extreme learning machine (IPSO-ELM) model was proposed and applied to the example of slope stability prediction.First in the basic particle swarm optimization (PSO), on the basis of which were liable to occur during the optimization process in order to overcome the local optimal problem, the introduction of the adaptive weight method, the improved particle swarm optimization (IPSO) for extreme learning machine (ELM) the weights of the input and hidden layer offset optimization, significantly improve the generalization ability and forecasting precision of the ELM model, and then will IPSO - ELM model, the PSO - ELM model and the predicted values and the real value of the ELM model comparison and analysis, the results show that IPSO - ELM model prediction approach in the real value, have higher prediction accuracy,The feasibility and effectiveness of IPSO-ELM model in slope stability evaluation are verified.
文章编号:     中图分类号:TD76    文献标志码:
基金项目:辽宁省教育厅重点项目(编号:601009877-36); 校青年(编号2020QN10)
引用文本:
赵允坤,胡 军,杨 斌.基于IPSO-ELM的边坡稳定性分析[J].有色金属工程,2022,(1):.

我们一直在努力打
造,精品期刊,传
播学术成果

全国咨询服务热线
86-10-63299224

杂志信息

期刊简介

相关下载

联系我们

电话:86-10-63299224/63299153

传真:010-63299754

Email:ysjsgc@bgrimm.com

邮编:100160

地址:北京市南四环西路188号总部基地十八区23号楼

关注微信公众号