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


Specify a method to gather training data

Topics

This keyword is related to the topics:

Specification

Alias: none

Argument(s): STRING

Default: no design of experiments data

Required/Optional Description of Group Dakota Keyword Dakota Keyword Description
Optional auto_refinement

Experimental auto-refinement of surrogate model

Description

The number of training points and the sources are specified on global, as well as the number of new training points required.

New training points are gathered by running the "truth" model using the method specified by dace_method_pointer. The DACE method will only be invoked if it has new samples to perform, and if new samples are required and no DACE iterator has been provided, an error will result.

The dace_method_pointer points to design of experiments method block used to generate truth model data.

Permissible methods include: Monte Carlo (random) sampling, Latin hypercube sampling, orthogonal array sampling, central composite design sampling, and Box-Behnken sampling.

Note that the number of samples specified in the method block may be overwritten, if the requested number of samples is less than minimum_points.