The DOEdriver, which is a driver that executes a Design of Experiments, includes a mandatory socket that takes a DOEgenerator. This DOEgenerator contains the algorithm that is used to select the design points that will be exectued by the DOEdriver. OpenMDAO includes several kinds of DOEdrivers which are listed below.
The FullFactorial DOEgenerator implements a full factorial Design of Experiments. In other words, it generates a set of design points that fully span the range of the parameters at the requested resolution.
(See the source documentation for more information on FullFactorial.)
The OptLatinHypercube component implements an algorithm that produces an optimal Latin hypercube based on an evolutionary optimization of its Morris-Mitchell sampling criterion.
(See the source documentation for more information on OptLatinHypercube.)