Comparison of the function regression method and data classification method for limit state function approximation

摘要

To reduce the computational burden of the reliability analysis of complex engineering application, approximate method is always used to construct the surrogate model of the implicit limit state function. Since the limit state function is a classifier of the failure domain and safe domain, its approximation can be established by the function regression method and data classification method. In this paper, these two methods are tested to several limit state functions including linear function, highly nonlinear function, high dimensional function, series system and parallel system. Least squares support vector machines are used to construct the surrogate models. A detail comparison of function regression method and data classification method for limit state function approximation is given. The conclusions of this paper can give guidance for the engineers to choose an appropriate approximate method in the engineering applications.