About Constraints

Normally, optimization tasks have some restrictions. These restrictions limit the values of the design responses or linear combinations of the design responses. If the constraint is not fulfilled the optimization result is not feasible.

Related Topics
In Other Guides
CONSTRAINT

SIMULIA Tosca Structure allows inequality constraints in all sensitivity-based algorithms. Equality constraints are only allowed in the controller-based approaches.

This means that the item EQ_VALUE defining the equality value may only be used for controller-based optimization. LE_VALUE and GE_VALUE are to be used for the upper and lower values of constraints in sensitivity-based optimization algorithms.

An equality constraint is given as:

Ψ=φ

where φ is the value of the design response. Inequality constraints are given as:

Ψ≤φ

Ψ≥φ

The MAGNITUDE can be defined as ABS or REL, short for absolute or relative value of the design response in the constraint. When using the relative value the design response is normalized with respect to the initial value of the design response (design response value of optimization iteration 0).