Page 1 of 1

Marginal values with BARON solver

Posted: Mon Mar 02, 2020 6:10 pm
by djdmlwns
Hi, I am a Ph.D. student in chemical engineering at UW-Madison.
I tried to obtain marginal values with BARON for a simple NLP problem, and they are not consistent with those from other nonlinear solvers such as IPOPT, CONOPT, etc.
Furthermore, BARON does not report the marginal values as shown in the screenshot below.
The tested problem is as follows.
Max. 2x1**2 + x2
s.t. 1-x1-x2>=0 (Eq1)
1-x1+x2>=0 (Eq2)
1+x1-x2>=0 (Eq3)
1+x1+x2>=0 (Eq4)

Clearly, the global optimal solution is (x1,x2) = (1,0).
For this problem, BARON does not report marginal values as shown in the above screenshot, while CONOPT gives correct marginal values (eq1 : -2.5, eq2: -1.5, and 0 for others).
I attached GAMS code for the problem for easier reproduction.
(365 Bytes) Downloaded 77 times
Does anyone know what is the problem here?
Thank you very much in advance!