1Blurb::
2Activate adaptive procedure for determining best rank representation
3
4Description::
5The adaptive algorithm proceeds as follows:
6-# Start from rank \c start_rank and form an approximation
7-# Adapt the current approximation by searching for a solution with lower rank that achieves L2 accuracy within epsilon tolerance of the reference.
8-# If a lower rank solution is found with comparable accuracy, then stop.  If not, increase the rank by an amount specified by \c kick_rank.
9-# Return to step 2 and continue until either \c max_rank is reached or a converged rank (rank less than current reference with comparable accuracy) is found.
10
11<b> Default Behavior </b>
12
13No cross validation for rank.
14
15Examples::
16This example shows specification of a rank adaptation starting at rank 2,
17incrementing by 2, and limited at rank 10.
18\verbatim
19model,
20	id_model = 'FT'
21	surrogate global function_train
22	  start_order = 5
23	  adapt_rank  start_rank = 2  kick_rank = 2  max_rank = 10
24	  solver_tolerance   = 1e-12
25	  rounding_tolerance = 1e-12
26	dace_method_pointer = 'SAMPLING'
27\endverbatim
28
29Note that \c adapt_rank and \c adapt_order can either be combined or
30used separately.
31
32See_Also::	model-surrogate-global-function_train-start_rank, model-surrogate-global-function_train-kick_rank, model-surrogate-global-function_train-max_rank
33