![]() |
Dakota Reference Manual
Version 6.16
Explore and Predict with Confidence
|
(Default) Use internal Dakota finite differences algorithm
Alias: none
Argument(s): none
Default: relative
Child Keywords:
| Required/Optional | Description of Group | Dakota Keyword | Dakota Keyword Description | |
|---|---|---|---|---|
| Optional | ignore_bounds | Do not respect bounds when computing gradients or Hessians | ||
| Optional (Choose One) | Step Scaling (Group 1) | relative | (Default) Scale step size by the parameter value | |
| absolute | Do not scale step-size | |||
| bounds | Scale step-size by the domain of the parameter | |||
The dakota routine is the default since it can execute in parallel and exploit the concurrency in finite difference evaluations (see Exploiting Parallelism in the Users Manual [4]).
When the method_source is dakota, the user may also specify the type of scaling desired when determining the finite difference step size. The choices are absolute, bounds, and relative. For absolute, the step size will be applied as is. For bounds, it will be scaled by the range of each parameter. For relative, it will be scaled by the parameter value.