Dakota Reference Manual  Version 6.4
Large-Scale Engineering Optimization and Uncertainty Analysis
 All Pages
sobol


Estimate dimension preference for automated refinement of stochastic expansion using total Sobol' sensitivity indices.

Specification

Alias: none

Argument(s): none

Default: generalized

Description

Determine dimension preference for refinement of a stochastic expansion from the total Sobol' sensitivity indices obtained from global sensitivity analysis. High indices indicate high importance for resolving statistical quantities of interest and therefore result in high dimension preference.

Examples

method,
        polynomial_chaos                        
          sparse_grid_level = 3
          dimension_adaptive p_refinement sobol 
            max_iterations  = 20                
            convergence_tol = 1.e-4