基于进化支持向量机的滑动面参数识别

Parameters recognition of slide surface using genetic-support vector machine

摘要: 提出了反算滑带岩土强度参数的一种新方法——进化支持向量机方法,该方法基于极限平衡分析方法计算出不同c、值下的安全系数与最危险滑动面,以这些数据作为学习样本,通过支持向量机学习,建立他们之间的非线性映射关系,然后基于这种关系,利用遗传算法搜索滑带岩土强度参数。结果表明,识别出的c、值可以满足工程的要求。

Abstract: A new method to back-calculate the strength parameters of landslide using genetic-support vector machine was proposed.Critical slip surfaces and minimum factors of safety were determined with different combinations of c and  using limited equilibrium analysis.The nonlinear relationship between c, and minimum factors of safety were established using support vector machine.The strength parameters of landslide were recognized based on this relationship.It was shown that the c and  values were sufficiently accurate for engineering applications.