coliny_solis_wets
Simple greedy local search method
Topics
package_scolib, package_coliny
Specification
Alias: None
Arguments: None
Child Keywords:
Required/Optional |
Description of Group |
Dakota Keyword |
Dakota Keyword Description |
---|---|---|---|
Optional |
The number of unsuccessful cycles prior to contraction. |
||
Optional |
Don’t allow expansion of the search pattern |
||
Optional |
Set the factor by which a search pattern can be expanded |
||
Optional |
Use a simple weighted penalty to manage feasibility |
||
Optional |
Amount by which step length is rescaled |
||
Optional |
Multiplier for the penalty function |
||
Optional |
Initial step size for derivative-free optimizers |
||
Optional |
Step length-based stopping criteria for derivative-free optimizers |
||
Optional |
Stopping criteria based on objective function value |
||
Optional |
Seed of the random number generator |
||
Optional |
Show algorithm parameters not exposed in Dakota input |
||
Optional |
Set method options not available through Dakota spec |
||
Optional |
Number of iterations allowed for optimizers and adaptive UQ methods |
||
Optional |
Stopping criterion based on objective function or statistics convergence |
||
Optional |
Number of function evaluations allowed for optimizers |
||
Optional |
Turn on scaling for variables, responses, and constraints |
||
Optional |
Identifier for model block to be used by a method |
Description
The Solis-Wets method is a simple greedy local search heuristic for continuous parameter spaces. Solis-Wets generates trial points using a multivariate normal distribution, and unsuccessful trial points are reflected about the current point to find a descent direction.
Note
See the page Package: SCOLIB for important information regarding all SCOLIB methods.
coliny_solis_wets
is inherently serial, no concurrency is used.
These specifications have the same meaning as corresponding
specifications for coliny_pattern_search
. Please see that
page for specification details.
In particular, coliny_solis_wets
supports dynamic rescaling of the
step length, and dynamic rescaling of the constraint penalty. The only
new specification is contract_after_failure
, which specifies the
number of unsuccessful cycles which must occur with a specific delta
prior to contraction of the delta.
Expected HDF5 Output
If Dakota was built with HDF5 support and run with the
hdf5
keyword, this method
writes the following results to HDF5:
Best Objective Functions (when
objective_functions
) are specified)Calibration (when
calibration_terms
are specified)