A randomized orthogonal array-based procedure for the estimation of first- and second-order Sobol' indices

In variance-based sensitivity analysis, the method of Sobol' [Sensitivity analysis for nonlinear mathematical models. Math Model Comput Exp. 1993;1:407–414] allows one to compute Sobol' indices (SI) using the Monte Carlo integration. One of the main drawbacks of this approach is that estimating SI requires a number of simulations which is dependent on the dimension of the model of interest. For example, estimating all the first- or second-order SI of a d-dimensional function basically requires or independent input vectors, respectively. Some interesting combinatorial results have been introduced to weaken this defect, in particular by Saltelli [Making best use of model evaluations to compute sensitivity indices. Comput Phys Commun. 2002;145:280–297] and more recently by Owen [Variance components and generalized Sobol' indices. SIAM/ASA J Uncertain Quantification. 2013;1:19–41], but the quantities they estimate still depend linearly on the dimension d. In this paper, we introduce a new approach to estimate all the first- and second-order SI by using only two input vectors. We establish theoretical properties of such a method for the estimation of first-order SI and discuss the generalization to higher order indices. In particular, we prove on numerical examples that this procedure is tractable and competitive for the estimation of all first- and second-order SI. As an illustration, we propose to apply this new approach to a marine ecosystem model of the Ligurian sea (northwestern Mediterranean) in order to study the relative importance of its several parameters. The calibration process of this kind of chemical simulators is well known to be quite intricate, and a rigorous and robust – that is, valid without strong regularity assumptions – sensitivity analysis, as the method of Sobol' provides, could be of great help. This article has supplementary material online.

[1]  Art B. Owen.  N A ] 8 M ay 2 01 2 Variance components and generalized Sobol ’ indices Art B , 2014 .

[2]  Max D. Morris,et al.  Sampling plans based on balanced incomplete block designs for evaluating the importance of computer model inputs , 2006 .

[3]  Saltelli Andrea,et al.  Sensitivity Analysis for Nonlinear Mathematical Models. Numerical ExperienceSensitivity Analysis for Nonlinear Mathematical Models. Numerical Experience , 1995 .

[4]  Harald Niederreiter,et al.  Random number generation and Quasi-Monte Carlo methods , 1992, CBMS-NSF regional conference series in applied mathematics.

[5]  Ronald L. Iman.  Latin Hypercube Sampling , 2008 .

[6]  B. Efron,et al.  The Jackknife Estimate of Variance , 1981 .

[7]  Thierry Alex Mara,et al.  Comparison of some efficient methods to evaluate the main effect of computer model factors , 2008 .

[8]  Wei-Liem Loh,et al.  A combinatorial central limit theorem for randomized orthogonal array sampling designs , 1996 .

[9]  Michael D. McKay,et al.  Evaluating Prediction Uncertainty , 1995 .

[10]  A. Owen.  Monte Carlo Variance of Scrambled Net Quadrature , 1997 .

[11]  C. Prieur,et al.  CERTIFIED REDUCED-BASIS SOLUTIONS OF VISCOUS BURGERS EQUATION PARAMETRIZED BY INITIAL AND BOUNDARY VALUES , 2010, 1010.1761.

[12]  T. Ishigami,et al.  An importance quantification technique in uncertainty analysis for computer models , 1990, [1990] Proceedings. First International Symposium on Uncertainty Modeling and Analysis.

[13]  H. Niederreiter,et al.  Monte Carlo and Quasi-Monte Carlo Methods in Scientific Computing , 1995 .

[14]  C. Prieur,et al.  Variance-based sensitivity analysis using harmonic analysis , 2012 .

[15]  D. Makowski,et al.  Chapter 3 Uncertainty and sensitivity analysis for crop models , 2006 .

[16]  Joseph A. C. Delaney.  Sensitivity analysis , 2018, The African Continental Free Trade Area: Economic and Distributional Effects.

[17]  Ruichen Jin,et al.  Analytical Variance-Based Global Sensitivity Analysis in Simulation-Based Design Under Uncertainty , 2005, DAC 2004.

[18]  Saltelli Andrea,et al.  Global Sensitivity Analysis: The Primer , 2008 .

[19]  I. Sobol,et al.  About the use of rank transformation in sensitivity analysis of model output , 1995 .

[20]  Peter Z. G. Qian.  Nested Latin hypercube designs , 2009 .

[21]  C. Lemieux.  Monte Carlo and Quasi-Monte Carlo Sampling , 2009 .

[22]  Paola Annoni,et al.  Variance based sensitivity analysis of model output. Design and estimator for the total sensitivity index , 2010, Comput. Phys. Commun..

[23]  A. O'Hagan,et al.  Probabilistic sensitivity analysis of complex models: a Bayesian approach , 2004 .

[24]  Charalambos A. Charalambides,et al.  Enumerative combinatorics , 2018, SIGA.

[25]  Stefano Tarantola,et al.  Random balance designs for the estimation of first order global sensitivity indices , 2006, Reliab. Eng. Syst. Saf..

[26]  Max D. Morris,et al.  Using Orthogonal Arrays in the Sensitivity Analysis of Computer Models , 2008, Technometrics.

[27]  A. Owen.  Scrambled net variance for integrals of smooth functions , 1997 .

[28]  Art B. Owen,et al.  Better estimation of small sobol' sensitivity indices , 2012, TOMC.

[29]  Bruno Sudret,et al.  Global sensitivity analysis using polynomial chaos expansions , 2008, Reliab. Eng. Syst. Saf..

[30]  W. Hoeffding.  A Class of Statistics with Asymptotically Normal Distribution , 1948 .

[31]  L. Devroye.  Non-Uniform Random Variate Generation , 1986 .

[32]  Marilaure Grégoire,et al.  Revisited ecosystem model (MODECOGeL) of the Ligurian Sea: seasonal and interannual variability due to atmospheric forcing , 2002 .

[33]  A. Saltelli,et al.  Making best use of model evaluations to compute sensitivity indices , 2002 .

[34]  A multivariate central limit theorem for randomized orthogonal array sampling designs in computer experiments , 2007, 0708.0656.

[35]  T. Klein,et al.  Asymptotic normality and efficiency of two Sobol index estimators , 2013, 1303.6451.

[36]  A. Owen.  Randomly Permuted (t,m,s)-Nets and (t, s)-Sequences , 1995 .

[37]  Geneviève Lacroix,et al.  Influence of meteorological variability on primary production dynamics in the Ligurian Sea (NW Mediterranean Sea) with a 1D hydrodynamic/biological model , 1998 .

[38]  K.,et al.  Nonlinear sensitivity analysis of multiparameter model systems , 1977 .