Dakota Reference Manual  Version 6.15
Explore and Predict with Confidence
 All Pages
max_hifi_evaluations


Maximum number of high-fidelity model runs to be used

Specification

Alias: none

Argument(s): INTEGER

Description

There are three stopping criteria by which the Bayesian experimental design algorithm will be terminated:

  • The high-fidelity model has been run max_hifi_evaluations times
  • The set of candidate design points have been exhausted
  • The relative change in the mutual information from one iteration to the next is sufficiently small The first criterion is controlled by this keyword specification

Default Behavior

If no maximum number of high-fidelity model runs is specified, only the last two stopping criteria listed above will be evaluated.

Usage Tips

Be wary that max_hifi_evaluations does not include any high-fidelity model evaluations that need to be run to produce the data set for the initial Bayesian calibration. This number only includes those evaluations performed on the sequentially chosen optimal design points

If max_hifi_evaluations is set to zero, the a single optimal experimental design point will be selected, and Dakota will exit without running the high- fidelity model.