The Dakota software provides workflows for a variety of iterative analyses, enabling the use of most any computational model for performing design optimization, model calibration, uncertainty quantification (UQ), and parametric/sensitivity analysis. In particular, its iterative systems analysis methods include:

• optimization with local and global, gradient- and nongradient-based methods;

• uncertainty quantification with sampling, reliability, stochastic expansion, and epistemic methods;

• model calibration using nonlinear least squares (deterministic) or Bayesian inference (stochastic); and

• parametric/sensitivity/variance analysis with design of experiments and parameter study methods.

These capabilities may be used on their own or as components within advanced solution strategies that coordinate multiple computational models and iteration methods, such as surrogate-based approaches to optimization, UQ and inference; mixed aleatory-epistemic UQ; and optimization under uncertainty.

Dakota is open source under GNU LGPL, with applications spanning defense programs for DOE and DOD, climate modeling, computational materials, nuclear power, renewable energy, and many others. It can be used as a stand-alone application or embedded as a library service, and supports multilevel parallelism using a hybrid scheduling model. Initiated in 1994, it strives to span the research to production spectrum, providing both a mature tool for production use as well as a foundation for new algorithm research.