Define, transform and sample random distributions


Key features:

  • Usual and custom marginals

  • Dependence modelling through copulas

  • Monte Carlo & Latin Hypercube sampling, low-discrepancy sequences


Connect your own simulation models to UQLab


Key features: 

  • Strings and inline function handles for analytical models

  • MATLAB m-files

  • Easy plugging of third party codes through wrappers

Compute fast surrogate models using polynomial chaos expansions


Key features: 

  • Full and sparse PC expansions

  • Quadrature, sparse grids, least-squares and least-angle regression

  • Advanced truncation schemes, custom basis specification

Compute robust surrogate models using Gaussian processes


Key features: 

  • Highly customizable trend and correlation functions

  • Maximum likelihood and cross-validation for estimating hyperparameters

  • Gradient-based and global optimizers

Identify the important input variables and their interactions


Key features:

  • Screening (Morris method)

  • Linear measures: Taylor series expansion (perturbation), standard regression coefficients

  • ANOVA: Sobol' indices through Monte Carlo and polynomial chaos expansions

Estimate probabilities of failure and distribution tails


Key features: 

  • FORM/SORM approximation methods

  • Sampling methods (Monte Carlo, importance sampling, subset simulation)

  • Kriging-based adaptive methods (AK-MCS)


Low Rank Approximations

Compute accurate surrogate models using canonical low rank polynomials 

Key features: 

  • Rank- and degree- adaptive approximation

  • Cross-validation-based basis construction

  • Customizable basis identification strategies

Polynomial-Chaos Kriging

Combine the global character of PCE with the local accuracy of Kriging​

Key features:

  • Sequential and optimal calculation strategies

  • Fully configurable Kriging and PCE sections

  • Sparse adaptive PCE trend estimation