Add a Control Variable Field node to define a control variable that varies as function of position within selected geometric entities (domains, boundaries, edges, or points). The control variable field is discretized using shape functions in the same way as other dependent variables in a multiphysics model. The discrete control variable degrees of freedom, on which the optimization solvers operate, represent values at element nodes. Right-click the node to add a Control Variable Bounds subnode.

Enter a Control variable name and Initial value.

Enter a Scale indicating a typical magnitude of the control variable. The relative solver tolerances refer to variables rescaled with respect to this scale, and it may also influences the search pattern of some optimization solvers.

This section contains settings for the shape functions used to discretize control variables. Select a Shape function type: Lagrange (the default) or Discontinuous Lagrange. Also select an Element order: Linear, Quadratic (the default), Cubic, Quartic, Quintic, Sextic, or Septic.

The default choice of Value type when using splitting of complex variables is Real. This means that if the solver is set up to split complex variables in real and imaginary parts, no imaginary part is allocated for the control variable field, which is therefore guaranteed to be real. Choose Complex to allocate both real and imaginary parts.