支持向量回归机在结构可靠性分析中的应用

 评估结构的可靠度常常遇到隐式功能函数，对于此种情况经典的结构可靠性分析方法难以实施。目前，响应面和人工神经网络等回归方法已经广泛地用来构建显式函数替代隐式功能函数进行结构可靠性分析。支持向量回归机是一种与人工神经网络类似的新回归方法，目前很少用于结构可靠性分析。鉴于支持向量回归机出色的小样本学习性能、良好的泛化性能，本文提出了基于支持向量回归机的两种适用于隐式功能函数的可靠性分析方法：基于支持向量回归机的Monte-Carlo模拟法和基于支持向量回归机的一次二阶矩法。算例表明支持向量回归机可以在抽样范围内很好的逼近真实的隐式功能函数，减少结构分析次数。失效概率计算结果与经典方法的对照说明所提方法的计算精度较高，具有一定的实用价值。

Support Vector Regression for Structural Reliability Analysis

Support vector regression (SVR),which is a novel re gr ession method and a potential alternative to artificial neural networks (ANN), d ue to its excellent learning capacity and generalization capability with a small amount of samples, has not been widely applied to structural reliability analys is currently.Thereby,two approaches based on SVR are proposed for structural re liability analysis,i.e.,SVR-based MCS and SVR-based FORM.In the proposed meth ods SVR is employed to approximate the implicit performance function.Examples wi th simple performance functions are used to demonstrate the application of SVR i n the structural reliability analysis.Some comparisons among the classical struc tural reliability analysis methods and the proposed approaches show that highly accurate approximation both in the performance functions and failure probabiliti es can be obtained by the SVR with a small amount of information.