multipoint_surrogate
Construct a surrogate from multiple existing training points
Specification
Alias: None
Arguments: None
JSON argument key: None
Child Keywords:
Required/Optional |
Description of Group |
JSON Group Key |
Dakota Keyword |
Dakota Keyword Description |
|---|---|---|---|---|
Required (Choose One) |
Type |
type |
Local multi-point model via two-point nonlinear approximation |
|
Multi-point surrogate approximation based on QMEA algorithm |
||||
Required |
Pointer to specify a “truth” model, from which to construct a surrogate |
|||
Optional |
Identifies the subset of the response functions by number that are to be approximated (the default is all functions). |
|||
Optional |
Give the model block an identifying name, in case of multiple model blocks |
|||
Optional |
Specify which variables block will be included with this model block |
|||
Optional |
Specify which reponses block will be used by this model block |
|||
Optional |
Enables hierarchical evaluation tagging |
Description
Multipoint approximations use data from previous design points to improve the accuracy of local approximations. The data often comes from the current and previous iterates of a minimization algorithm.
Currently, only the
Two-point Adaptive Nonlinearity Approximation (TANA-3) method of
[XG98] is supported with the tana keyword.
The
truth model to be used to generate the value/gradient data used in the
approximation is identified through the required
truth_model_pointer specification.

