Dakota 6.13

Released: November 16, 2020

Highlight: Dakota GUI Updates

  • Chartreuse
    • New "Sandbox View" for fast visualizations of generic data using Chartreuse.
    • Support added to Chartreuse for CSV files
    • Four-dimensional Chartreuse scatter plots (i.e. time-based node coloring)
  • Dakota Input File Editing
    • New form-based editors for Dakota interface blocks and hybrid method blocks.
    • Limited support for visualization of Dakota uncertainty variables (normal, lognormal, weibull)
    • Pre-processing markup supported in Dakota text editor, which also provides a new mechanism for assigning multiple NGW-based analysis drivers at runtime.
    • Dark theme support for Dakota text editor
  • QOI
    • New column-based QOI extractors for extracting fields from tabular/CSV-based data
    • Warning: The qoiExtractor node in Next-Gen Workflow has received backwards-incompatible changes. You must delete your qoiExtractor nodes and reconfigure them upon switching to 6.13
  • Misc.
    • "dprepro" node added to Next-Gen Workflow
    • General enhancements for the New Dakota Study wizard

Enabling / Accessing:  Dakota GUI ships with Dakota and is available for Windows, Mac, and RHEL7.

Documentation:  An enhanced version of the Dakota GUI manual now ships with the GUI, giving you easy access to a wealth of reference material for using the GUI.  The 6.13 GUI manual is also available here.

Highlight: Surrogate Models

New polynomial and Gaussian proces surrogate models notably gained save/load capability, metrics including cross-validation, and a Python interface and are more widely available in Dakota methods. Details:

  • Serialization together with Pybind11-based Python wrappers allow export of surrogates from Dakota for later use directly from Python.
  • New GP, including surrogate export is now available in EGO, EGRA, and global interval methods
  • New minor features: advanced options configuration via external XML file, none/reduced_quadratic trend option, GP performance improvements, verbosity control, MLE estimation history, embed variable/response labels

Enabling / Accessing: Most surrogate features are default-enabled in Dakota, however the Python interface must be manually enabled at compile time by specifying DAKOTA_PYTHON_SURROGATES:BOOL=TRUE.

Documentation: See Section 8.4.7 in the Dakota 6.13 User's Manual, together with reference manual documentation for global surrogate models and the methods mentioned above.