基于改进灰狼算法优化ELM的边坡稳定性评价  

Slope Stability Evaluation Based on ELM Optimized by Improved Grey Wolf Algorithm

作　　者：马艳梅[1] 毕晓茜[2] MA Yanmei;BI Xiaoqian(Huainan Vocational and Technical College,Huainan Anhui 232001,China;Xuzhou Institute of Technology,Xuzhou Jiangsu 221018,China)

机构地区：[1]淮南职业技术学院,安徽淮南232001 [2]徐州工程学院,江苏徐州221018

出　　处：《长春工程学院学报（自然科学版）》2023年第3期24-28,共5页Journal of Changchun Institute of Technology：Natural Sciences Edition

基　　金：安徽省自然科学基金项目(KJ2021A1582)。

摘　　要：采用改进的灰狼算法对极限学习机进行了优化,以改善边坡稳定性评估的准确性。首先,在GWO中加入了逆向和非线性的收敛因子,并给出了一种新的GWO优化方法;其次,为了改善ELM的性能,利用IGWO算法优化了ELM的输入层次权重和隐含层偏差,从而得到了IGWO-ELM的最佳模式。最后,选取重度、黏聚力、摩擦角、边坡角和坡高5个指标作为边坡稳定性评估参数,研究结果表明,与ELM、PSOELM、GWO-ELM相比,应用IGWO-ELM对边坡的稳定性进行评估,其结果更为准确。The improved Grey Wolf algorithm is used to optimize the extreme learning machine(ELM)to improve the accuracy of slope stability evaluation.It firstly makes the inverse and non linear convergence factors into the Grey Wolf Optimization algorithm(GWO)to propose a new improved GWO(IGWO).It u-ses IGWO algorithm to optimize the input level weights and implied layer deviations of the ELM then,so as to improve the performance of the ELM and obtain the best model of the IGWO-ELM.Finally,five inde-xes such as gravity,cohesion,friction angle,slope angle and slope height are chosen to be the parameters of slope stability evaluation.It is indicated that the result of the stability of slopes evaluated by IGWO-ELM is more accurate compared with ELM,PSO-ELM and GWO-ELM.

关 键 词：极限学习机 边坡稳定性 灰狼优化算法 反向学习 粒子群算法 

分 类 号：TU43[建筑科学—岩土工程] TD824.71[建筑科学—土工工程]