Adaptive Response Surface Method for Efficient Bayesian Reliability Based Design Optimization
Citation
Sadilingam, Gopi Krishna (2011). Adaptive Response Surface Method for Efficient Bayesian
Reliability Based Design Optimization. -- In Proceedings: 7th Annual Symposium: Graduate Research and Scholarly Projects. Wichita, KS: Wichita State University, p. 126-127
Abstract
To tackle engineering design problems engaging both aleatory and epistemic uncertainties, Reliability-Based Design
Optimization (RBDO) has been integrated with Bayes Theorem, referred to as Bayesian RBDO. However, Bayesian RBDO
becomes expensive when employing the First- or Second-Order Reliability Methods for reliability predictions. This paper
proposes an Adaptive Response Surface Method (ARSM) for efficient Bayesian reliability analysis and design optimization. The
ARSM integrates the iterative design optimization process with the local response surface methodology through an adaptive
sampling scheme. Through this integration, the information for reliability analysis generated at early design stages can be used
adaptively to construct local response surfaces for later design iterations. Thus, the computational efficiency of the Bayesian
RBDO can be improved as substantially fewer experiments are required in the overall design process. The proposed methodology
is demonstrated with a ground vehicle lower control arm design case study.
Description
Paper presented to the 7th Annual Symposium on Graduate Research and Scholarly Projects (GRASP) held at the Marcus Welcome Center, Wichita State University, May 4, 2011.
Research completed at the Department of Industrial and Manufacturing Engineering