An efficient dimension-adaptive uncertainty propagation approach

In recent years, there has been a growing interest in uncertainty propagation, and a wide variety of uncertainty propagation methods exist in literature. In this paper, an uncertainty propagation approach is developed by using high-dimensional model representation (HDMR) and dimension reduction (DR) method technique in the stochastic space to represent the model output as a finite hierarchical correlated function expansion in terms of the stochastic inputs starting from lower-order to higher-order component functions. To save the computational cost, a dimension-adaptive version of the additive decomposition is proposed to detect the important component functions to reduce the terms. The proposed method requires neither the calculation of partial derivatives of response, as in commonly used Taylor expansion/perturbation methods, nor the inversion of random matrices, as in the Neumann expansion method. Two numerical examples show the efficiency and accuracy of the proposed method.