DAKOTA, A Multilevel Parallel Object-Oriented Framework for Design Optimization, Parameter Estimation, Sensitivity Analysis, and Uncertainty Quantification

The DAKOTA (Design Analysis Kit for OpTimizAtion) iterator toolkit is a flexible, extensible interface between simulation codes and iterative systems analysis methods. The toolkit implements optimization with a variety of gradient and nongradient-based methods, uncertainty quantification with nondeterministic propagation methods, parameter estimation with nonlinear least squares solution methods, and sensitivity analysis with general-purpose parameter study capabilities. By employing object-oriented design to implement abstractions of the key concepts involved in iterative systems analyses, the DAKOTA toolkit provides a flexible and extensible problem-solving environment which uses point solutions from simulation codes for answering fundamental engineering questions, such as “what is the best design?”, “how safe is it?”, or “how much confidence do I have in my answer?”. In addition to these iterative systems analysis capabilities, advanced users can employ state of the art capabilities for (1) exploiting parallelism at multiple levels (coarse-grained and fine-grained) and (2) building cascaded, nested, concurrent, and/or adaptive strategies which utilize multiple iterators and models to enact hybridization, sequential approximation, stochastic optimization, or mixed continuous-discrete optimization. This report serves as a user’s guide and reference for the DAKOTA software and provides capability overviews, command specifications, and installation and execution instructions.