Optional Correlation Function Method in Kriging for Stochastic Optimal Design

Abstract:

This paper is focused on the three strategies to improve efficiency and accuracy of approximate optimization models using Kriging. The strategies are performed by the stochastic process which is called stochastic-localization method as the criterion to move the local domains and the design of experiments, the classical design and space-filling design. We also propose the methodology conducted by the max-min reused sampling and a sequential adaptation algorithm of correlation functions. The proposed strategies are applied to the known analytical function such as Sandgren’s pressure vessel and three-bar truss for practical examples.

• Periodical: Materials Science Forum (Volumes 580-582)
• Pages: 609-612
• DOI: https://doi.org/10.4028/www.scientific.net/MSF.580-582.609
• Online since: June 2008
• Authors: Jong Bin Im, Young Hee Ro, Soo Yong Lee, Jung Sun Park
• Keywords: Kriging, Max-Min Reused Sampling, Stochastic Localization Method (SLM)

References

This article has no references.