Efficient dependency models: Simulating dependent random variables

Abstract

Dependency functions of dependent variables are relevant for (i) performing uncertainty quantification and sensitivity analysis in presence of dependent variables and/or correlated variables, and (ii) simulating random dependent variables. In this paper, we mathematically derive practical dependency functions for classical multivariate distributions such as Dirichlet, elliptical distributions and independent uniform (resp. gamma and Gaussian) variables under constraints that are ready to be used. Since such dependency models are used for sampling random values and we have many dependency models for every joint cumulative distribution function, we provide a way for choosing the efficient sampling function using multivariate sensitivity analysis. We illustrate our approach by means of numerical simulations.