A Quadrature-Based Sampling Technique for Robust Design With Computer Models

摘要: Several methods have been proposed for estimating transmitted variance to enable robust parameter design using computer models. This paper presents an alternative technique based on Gaussian quadrature which requires only 2n+1 or 4n+1 samples (depending the accuracy desired) where n is number of randomly varying inputs. The quadrature-based assessed a hierarchical probability model. can estimate standard deviation within 5% in over 95% systems much better than Hammersley Sequence Sampling, Latin Hypercube and Quadrature Factorial Method under similar resource constraints. If most accurate existing method, afforded ten times samples, it provides approximately same degree as method. Two case studies confirmed main conclusions also suggest method becomes more robustness improvements are made.Copyright © 2005 by ASME