Dakota Reference Manual  Version 6.4
Large-Scale Engineering Optimization and Uncertainty Analysis
 All Pages

(Deprecated keyword) Augments an existing random sampling study


Alias: none

Argument(s): none


This keyword is deprecated. Instead specify sample_type random with refinement_samples.

An incremental random sampling approach will augment an existing random sampling study with refinement_samples to get better estimates of mean, variance, and percentiles. There is no constraint on the number of samples in the second or subsequent sets as there is with incremental LHS.

Typically, this approach is used when you have an initial study with sample size N1 and you want to perform an additional N2 samples. Ideally, a Dakota restart file containing the initial N1 samples, so only N2 (instead of N1 + N2) potentially expensive function evaluations will be performed.

Usage Tips

The incremental approach is useful if it is uncertain how many simulations can be completed within available time.


Suppose an initial study is conducted using sample_type random with samples = 50. A follow-on study uses a new input file where samples = 50, and refinement_samples = 10, resulting in 50 reused samples (from restart) and 10 new random samples. The 10 new samples will be combined with the 50 previous samples to generate a combined sample of size 60 for the analysis.

The method block for the incremented study input60.in would be the following:

    sample_type incremental_random
    samples = 50
      refinement_samples = 10

The syntax for running the second sample set night be:

dakota -i input60.in -r dakota.50.rst

where dakota.50.rst is the restart file containing the results of the previous study.