Variance-based sensitivity indices of computer models with dependant inputs: The Fourier amplitude sensitivity test

Several methods are proposed in the literature to perform global sensitivity analysis of computer models with independent inputs. Only a few allow for treating the case of dependent inputs. In the present work, we investigate how to compute variance-based sensitivity indices with the Fourier amplitude sensitivity test. This can be achieved with the help of the inverse Rosenblatt transformation or the inverse Nataf transformation. We illustrate this on two distinct benchmarks. As compared to the recent Monte Carlo based approaches recently proposed by the same authors [Mara, T.A., Tarantola, S., and Annoni, P., Non-parametric methods for global sensitivity analysis of model output with dependent inputs, Env. Model. Software, 72:173–183, 2015], the new approaches allow us to divide the computational effort by 2 to assess the entire set of first-order and total-order variance-based sensitivity indices.

TARANTOLA Stefano;  MARA Thierry; 

2017-12-21

BEGELL HOUSE INC

JRC109470

2152-5080,   

https://publications.jrc.ec.europa.eu/repository/handle/JRC109470,   

10.1615/Int.J.UncertaintyQuantification.2017020291,