How can I set bounds for ga that change with the values of the optimization variables?

Hi. I have lower and upper bounds for my ga optimization variables that depend on those variables' values. Is there a way to input the values of the variables in each iteration of the ga into functions that compute the lower and upper bounds vectors and input those lb and ub back into the ga?

Answer 1

[John D'Errico]

You cannot. Bounds are FIXED constants. They do not vary at all, and the optimizer assumes that as fact.

If you have a constraint that is a function of the parameters, then that is a nonlinear constraint. Or a linear one, if it can be writtten as a linear combination of the parameters.

[Radu Andrei Matei]

Thanks. I've tried to apply this and of course, a cascade of other problems appeared. I've posted another question relating this one, I'd greatly appreciate it if you could check that one out aswell. But in any case, thanks for your quick and useful answer :)

Answer 2

[Torsten]

Instead of using the array lb and ub, you will have to prescribe such bounds that depend on the solution variables in the Constraint Function, usually called "nonlcon" in the documentation.

[Walter Roberson]

However, it is common to use lb ub and linear or nonlinear constraints. lb and ub are the fastest kind of constraints; linear constraints are a little slower (they require multiplications), and nonlinear constraints are slowest.

[Radu Andrei Matei]

Hi, thank you for your answer. I've accepted the other one as it was similarly helpful but responded before.