A single-loop shifting vector method with conjugate gradient search for reliability-based design optimization
Year: 2020
Type: article
Abstract: Raktim Biswas & Deepak Sharma*Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, IndiaCONTACT Deepak Sharma [email protected] methods for reliability-based design optimization are characterized by high computational efficiency. However, the accuracy of these methods may be low for highly nonlinear probabilistic constraints. A new hybrid method called ‘SLShV-CG’ is proposed for improving accuracy in which the single-loop method is coupled with the shifting vector approach of the sequential optimization and reliability assessment method. This shifting vector approach is incorporated with the probabilistic constraints without approximating them for reliability analysis. The most probable points for the probabilistic constraints are found using the Karush–Kuhn–Tucker conditions and the conjugate gradient search direction is used for determining the approximate most probable point in every iteration. The proposed method is tested on four mathematical and four engineering reliability-based design optimization problems, and the accuracy of the solutions is verified using Monte Carlo simulations. The results demonstrate the better accuracy and computational efficiency of the proposed method over six reliability-based design optimization methods from the literature.
Source: Engineering Optimization
Authors: Raktim Biswas, Deepak Sharma
Institution: Indian Institute of Technology Guwahati
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W3033764742