# mixed

Maintain continuous/discrete variable distinction

**Specification**

*Alias:*None*Arguments:*None*Default:*relaxed (branch and bound), mixed (all other methods)

**Description**

The variables domain specifies how the discrete variables are treated.
If the user specifies `mixed`

in the variable specification block,
the continuous and discrete variables are treated separately. If the
user specifies `relaxed`

in the variable specification block, the
discrete variables are relaxed and treated as continuous
variables. This may be useful in optimization problems involving both
continuous and discrete variables when a user would like to use an
optimization method that is designed for continuous variable
optimization. All Dakota methods have a default value of mixed for
the domain type except for the branch-and-bound method which has a
default domain type of relaxed. Note that the branch-and-bound method
is under development at this time. Finally, note that the domain
selection applies to all variable types: design, aleatory uncertain,
epistemic uncertain, and state.

With respect to domain type, if the user does not specify an
explicit override of `mixed`

or `relaxed`

, Dakota infers
the domain type from the method. As mentioned above,
all methods currently use a mixed domain as a default, except
the branch-and-bound method which is under development.