[Ipopt] Question regarding solving a (possibly?) non-convex model to global optimality

Juan Galvez jjgalvez at um.es
Sat Nov 26 16:56:36 EST 2011


I don't know if this is the right place to ask this (if not, please  
point me somewhere I can ask this question).

I need to solve a nonlinear programming model to global optimality  
using ipopt. I'm fairly new to this field so my knowledge is rather  

What I want to do is minimize a convex objective function subject to  
linear constraints and one constraint which I believe is non-convex.
This constraint is of the following form:

sum_{k}{a(k)*g(k)} <= C

where a(k) is a continuous variable in [0,1], g(k) is a continuous  
variable in [0,X] and C is a constant.

I can solve the problem with ipopt but I understand that to guarantee  
global optimality the problem has to be convex. I'm not sure if the  
above constraint makes the problem non-convex (and therefore ipopt  
cannot guarantee global optimality), and if so, is it possible to  
transform this constraint to equivalent convex constraints?

I'd appreciate any help.

Best regards,
Juan Galvez

More information about the Ipopt mailing list