Dakota Reference Manual  Version 6.2
Large-Scale Engineering Optimization and Uncertainty Analysis
 All Pages

(Experimental) efficient subspace method (ESM)


This keyword is related to the topics:


Alias: nond_efficient_subspace

Argument(s): none

Required/Optional Description of Group Dakota Keyword Dakota Keyword Description
Optional emulator_samples Number of data points used to train the surrogate model or emulator
Optional batch_size The number of points to add in each batch.
Optional distribution

Selection of cumulative or complementary cumulative functions

Optional probability_levels Specify probability levels at which to estimate the corresponding response value
Optional gen_reliability_levels Specify generalized relability levels at which to estimate the corresponding response value
Optional rng

Selection of a random number generator

Optional samples

Number of samples for sampling-based methods

Optional seed

Seed of the random number generator

Optional model_pointer

Identifier for model block to be used by a method


ESM is experimental and its implementation is incomplete. It is an active subspace method, intended for use with models with high dimensional input parameter spaces and analytic gradients. The method works by evaluating the response gradient at a number of points in the input parameter space and using a singular value decomposition to identify key linear combinations of input directions along which the response varies. Then UQ is performed in the reduced input parameter space.