Specifies design optimization variables.


AcuSolve Command


DESIGN_VARIABLE {parameters}


This command has no qualifier.


initial_value (real) [=0.0]
Initial value assigned to the design variable.
lower_bound (real) [=-1.0]
Lower bound of design variable.
upper_bound (real) [=1.0]
Upper bound of design variable.
max_update_factor (real) [=0.05]
Max relative step change allowed for the design variable to change between iterations. The max update factor is relative to the upper_bound and the lower_bound difference.
supervised_training (array) [={1.0}]
Manual supervised training is used as input data to create a so called response surface. The corresponding response function values are computed by AcuSolve for the given input training data. The response surface is constructed using a Gaussian Process, which essentially is a function approximating the response values as a function of the design variables.


The shape of the deformed geometry is controlled by design variable commands. For each variable a specified shape is defined, for example using mesh displacements defined in a nodal boundary condition command. Design variables may also control parameters by using multiplier functions of type set to design_variable.
	initial_value            = 0.0
	lower_bound              = -1.0
	upper_bound              = 1.0
	max_update_factor        = 0.05
	supervised_training      = {0.1,0.2,0.5,…. }