基于蚁群优化最小二乘支持向量机模型的边坡稳定性分析

SLOPE STABILITY ANALYSIS USING LEAST SQUARE SUPPORT VECTOR MACHINE OPTIMIZED WITH ANT COLONY ALGORITHM

边坡稳定性与其影响因素之间存在着复杂的非线性关系。通过分析影响边坡稳定性的主要因素,采用支持向量机建立边坡稳定性和影响因素之间的非线性关系;同时,考虑到支持向量机参数对预测效果的影响,采用连续蚁群算法对其进行优化选择,从而提出边坡稳定性预测的蚁群优化支持向量机模型。锦屏一级右岸拱肩槽部位谷坡为顺向坡,绝大部分基岩裸露,自然边坡为大理岩边坡,现状稳定。结合锦屏一级右岸拱肩槽边坡,采用蚁群优化支持向量机模型对其稳定性进行预测分析,预测结果与实际情况吻合较好,说明蚁群优化支持向量机模型在边坡稳定性分析中具有良好的实际应用价值。

The relationship between slope stability and influencing factors is complex and nonlinear. Support vector machine （ SVM） is used to build the nonlinear relationship. The parameters of SVM are optimized with a continu- ous ant colony algorithm （ CACA ）. Thus the ACAC - SVM is proposed for forecasting slope stability. Slope of right - bank spandrel groove at Jinping is forward slope. A majority of the slope surface is bare and comprised of bedrock. the natural slope is comprised of marbleand stability now. The CACA - SVM is used to analysis the stability of the slope of right - bank spandrel groove at Jinping. The results are in good agreement with the actual data, which indicates that the CACA - SVM can be well applied to the analysis of slope stability.