进化-最小二乘支持向量机的边坡稳定性估计  

Genetic-least square support vector machine estimation of slope stability

作　　者：马文涛[1] 孔亮[2] 

机构地区：[1]宁夏大学数学计算机学院,银川750021 [2]青岛理工大学理学院,青岛266033

出　　处：《岩土力学》2009年第12期3876-3880,共5页Rock and Soil Mechanics

基　　金：新世纪优秀人才支持计划资助项目(No.NCET-05-0895);宁夏自然科学基金项目(No.NZ0823);宁夏大学自然科学基金(No.ZR200702)

摘　　要：针对最小二乘支持向量机的参数选择问题,用遗传算法来搜索最小二乘支持向量机的相关参数,避免了人工搜索参数的盲目性,提高了模型的推广性能。根据大量的实际边坡工程数据,建立了基于进化-最小二乘支持向量机的边坡稳定性模型,并将其应用于估计丁家河磷矿自然边坡稳定状况。计算结果与工程实际情况一致,表明了该方法的有效性和合理性。The genetic algorithm is used to search the parameters of least square support vector machine in order to avoid the blindness of manual searching and improve the generalization of the model. Based on the large amount of measured data of practical slope engineering, a genetic-least square support vector machine estimation model of slope stability is set up; and then it is applied to predict the stability factor of natural slope in Dingjiahe phosphorus mine. The agreement of the theoretical results with the actual situation of natural slope shows that the proposed model is effective and reliable.

关 键 词：边坡稳定 最小二乘支持向量机 遗传算法 参数选择 

分 类 号：TU457[建筑科学—岩土工程]