scalarization
Fit MLMC sample allocation to a mixture of terms of means and standard deviations.
Specification
Alias: None
Arguments: None
JSON argument key: None
Child Keywords:
Required/Optional |
Description of Group |
JSON Group Key |
Dakota Keyword |
Dakota Keyword Description |
|---|---|---|---|---|
Optional |
Coefficients for linear scalarization (combination) of responses |
|||
Optional |
Solve the optimization problem for the sample allocation by numerical optimization in the case of sampling estimator targeting the scalarization. |
Description
Fit MLMC sample allocation to control the variance of the estimator for a mixture of terms of means and standard deviations. The exact scalarized formulation is given by the keyword scalarization_response_mapping.
Examples
The following method block
method,
model_pointer = 'HIERARCH'
multilevel_sampling
pilot_samples = 20 seed = 1237
convergence_tolerance = .01
allocation_target = scalarization
scalarization_response_mapping = 1 0 0 0
0 0 1 3
{
"method": [
{
"multilevel_sampling": {
"model_pointer": "HIERARCH",
"pilot_samples": [20],
"seed_sequence": [1237],
"convergence_tolerance": {
"value": 0.01
},
"allocation_target": {
"scalarization": {
"scalarization_response_mapping": [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 3.0]
}
}
}
}
]
}
uses the standard_deviation as sample allocation target by computing its variance. In this example, we assume a problem with two responses where the first line in scalarization_response_mapping refers to the first response, the second line to the second response. In the first line we only use 1 times the mean as quantity of interest. For the second response, we use 1 time the mean plus 3 times the standard devitation of the second quantity of interested. This behavior mimics the keywords primary_response_mapping and secondary_response_mapping.

