An adaptive dimension decomposition and reselection method for reliability analysis

Authors:
Chao Hu;Byeng D. Youn;Heonjun Yoon
Affiliations:
Department of Mechanical Engineering, The University of Maryland at College Park, College Park, USA 20742;School of Mechanical and Aerospace Engineering, The Seoul National University, Seoul, Republic of Korea;School of Mechanical and Aerospace Engineering, The Seoul National University, Seoul, Republic of Korea
Venue:
Structural and Multidisciplinary Optimization
Year:
2013

Citing 14
Cited 0

Stochastic finite elements: a spectral approach

Stochastic finite elements: a spectral approach
Adaptive sparse grid multilevel methods for elliptic PDEs based on finite differences

Computing
Simulation and the Monte Carlo Method

Simulation and the Monte Carlo Method
The Wiener--Askey Polynomial Chaos for Stochastic Differential Equations

SIAM Journal on Scientific Computing
Dimension-adaptive tensor-product quadrature

Computing
High-Order Collocation Methods for Differential Equations with Random Inputs

SIAM Journal on Scientific Computing
Multi-Element Generalized Polynomial Chaos for Arbitrary Probability Measures

SIAM Journal on Scientific Computing
Sparse grid collocation schemes for stochastic natural convection problems

Journal of Computational Physics
A Sparse Grid Stochastic Collocation Method for Partial Differential Equations with Random Input Data

SIAM Journal on Numerical Analysis
The multi-element probabilistic collocation method (ME-PCM): Error analysis and applications

Journal of Computational Physics
An adaptive hierarchical sparse grid collocation algorithm for the solution of stochastic differential equations

Journal of Computational Physics
Multi-element probabilistic collocation method in high dimensions

Journal of Computational Physics
An adaptive high-dimensional stochastic model representation technique for the solution of stochastic partial differential equations

Journal of Computational Physics
Dimension-wise integration of high-dimensional functions with applications to finance

Journal of Complexity

Quantified Score

Hi-index	0.00

Visualization

Abstract

Recently, the research community in reliability analysis has seen a strong surge of interest in the dimension decomposition approach, which typically decomposes a multi-dimensional system response into a finite set of low-order component functions for more efficient reliability analysis. However, commonly used dimension decomposition methods suffer from two limitations. Firstly, it is often difficult or impractical to predetermine the decomposition level to achieve sufficient accuracy. Secondly, without an adaptive decomposition scheme, these methods may unnecessarily assign sample points to unimportant component functions. This paper presents an adaptive dimension decomposition and reselection (ADDR) method to resolve the difficulties of existing dimension decomposition methods for reliability analysis. The proposed method consists of three major components: (i) an adaptive dimension decomposition and reselection scheme to automatically detect the potentially important component functions and adaptively reselect the truly important ones, (ii) a test error indicator to quantify the importance of potentially important component functions for dimension reselection, and (iii) an integration of the newly developed asymmetric dimension-adaptive tensor-product (ADATP) method into the adaptive scheme to approximate the reselected component functions. The merits of the proposed method for reliability analysis are three-fold: (a) automatically detecting and adaptively representing important component functions, (b) greatly alleviating the curse of dimensionality, and (c) no need of response sensitivities. Several mathematical and engineering high-dimensional problems are used to demonstrate the effectiveness of the ADDR method.