The Distributed Design and Analysis of Computer Experiments (DDACE) package is a valuable tool for analysts and engineers who want to:
Additionally, DDACE can generate parameter-dependent sensitivity information for use in optimization algorithms. In particular, DDACE sensitivity information was used to compute finite-difference gradients for OPT++’s nonlinear interior-point method, which resulted in an objective function value decrease for a thermal analysis parameter identification problem arising from an extreme ultraviolet lamp model. Within the DAKOTA framework, DDACE can be combined with surrogate-based optimization and optimization under uncertainty methods. The package contains a variety of sampling methods, including:
For more information, please visit the DDACE web site.