Robust design optimization has gained increasing concern in the engineering design process because it can provide an economical design that is insensitive to variations in the input variables without eliminating their causes. Robustness assessment, which estimates the mean values and standard deviations of the objective function and constraint functions, is an important and inevitable component in robust design optimization. A hybrid dimension-reduction method is proposed for efficient and accurate robustness assessment in this paper. It is a combination of the univariate dimension-reduction method and the bivariate dimension-reduction method for the balance of efficiency and accuracy. The significant random variables are identified by their variation contributions to the output variation of the objective function and constraints using the univariate dimension-reduction method. The approximate part of the significant random variables in the approximate function is then extended to use the bivariate dimension-reduction method, while the rest of the approximate part is still kept using the univariate dimension-reduction method. It then couples with a numerical optimization algorithm, formulating an iterative robust design-optimization procedure. Two engineering examples are used to illustrate its advantages. The proposed method is found to be competitive in estimating the mean values and standard deviations of the objective function and constraint functions in robust design optimization.
Hybrid Dimension-Reduction Method for Robust Design Optimization
AIAA Journal ; 51 , 1 ; 138-144
2012-11-26
7 pages
Article (Journal)
Electronic Resource
English
Dimension Reduction for Aerodynamic Design Optimization
AIAA | 2011
|Bayesian Reliability-Based Design Optimization using Eigenvector Dimension Reduction ( EDR ) method
Automotive engineering | 2007
|System Reliability-Based Design Optimization Using MPP-Based Dimension Reduction Method
British Library Conference Proceedings | 2008
|Dimension Reduction in Direction Finding Optimization
IEEE | 2021
|