nonlinear optimization with dynamic constraints

Hi all

Is it possible to assign a dynamic set of constraints when using an optimization routine such as fmincon? Specifically I am trying to solve a problem with 3 variables and I'd like to have the bounds on the last variable depend on the sign of the first variable.

edit: the bounds on the third variable are set in advance (bound_pos & bound_neg).

Cheers

Jakob

Answer 1

[Alan Weiss]

While you can program anything you like into a nonlinear constraint function, I recommend that you solve the problem twice, one with bounds set assuming that x(1) is positive, the other assuming that x(1) is negative. Then examine your solutions and take the one that is consistent with your assumption.

Alan Weiss

MATLAB mathematical toolbox documentation

[Jakob Sievers]

Thanks for this comment as well. I figured out that I might be able to narrow down the sign of that third parameter before using an optimization routine so I could make due with just one run. Cheers