![]() |
Dakota Reference Manual
Version 6.15
Explore and Predict with Confidence
|
candidate generation by advancement of maximum rank
Alias: none
Argument(s): none
Within adaptive refinement processes, this specification selects the advancement of max_rank
to create new refinement candidates. This option requires the use of adapt_rank
, for which this advancement creates an expanded range for the cross validation (CV) process. Once the CV process saturates (the best rank is less than the max), then no further advancement can occur and the refinement process (for a particular model level/fidelity) halts.