Hi
I have a question:
BARON is said to be a global optimizer, but for some numerical examples in my model, metaheuristic algorithm produces better results than BARON.
It can help me a lot if someone can explain how it is possible.
Thanks.
\
BARON does global optimization, but there is a parameter optcr which you should know about. This parameter sets the stopping criteria for the relative optimality gap, so optcr = 0.01 basically means the solver can stop if the relative gap to global optimality is guaranteed to be less than 1% (other stopping criteria permitting, of course). optcr is not 0 by default, so make sure you check the status after a solve as it will say whether the solution at termination is global optimal, locally optimal, etc. To insist on global optimality, try using the command option optcr=0; but keep in mind that you might still get a suboptimal solution if it stops for another reason like reaching the time limit or iteration limit.
About optcr: http://www.gams.com/mccarl/mccarlhtml/optcr.htm
About solver statuses (see modelStat): http://www.gams.com/mccarl/mccarlhtml/model_attributes_mainly_used_a.htm (status 1 means global optimality)
I hope this helps.
-Trish
On Saturday, July 19, 2014 6:39:45 AM UTC-4, Sanaz Tayefeh Hashemi wrote:
Hi
I have a question:
BARON is said to be a global optimizer, but for some numerical examples in my model, metaheuristic algorithm produces better results than BARON.
It can help me a lot if someone can explain how it is possible.
Thanks.
–
To unsubscribe from this group and stop receiving emails from it, send an email to gamsworld+unsubscribe@googlegroups.com.
To post to this group, send email to gamsworld@googlegroups.com.
Visit this group at http://groups.google.com/group/gamsworld.
For more options, visit https://groups.google.com/d/optout.