Page 1 of 1

MIPTRACE (objective value not in the exponential format)

Posted: Mon Apr 22, 2019 7:13 pm
by nuno_r
Hello,
I would like to trace the objective values of the best solutions found across the several iterations of my model.
For this purpose, I have used the miptrace option.
The problem is that the objective value of my problem is very big, higher than 1000000000000000, and the miptrace option rounds this value presenting it in the exponential format.
I need to see the entire number. Do you know if this is possible?
If not, is it possible to save in a gdx all the solutions found by GAMS over the iterations?

Example:
1, S, 0, 0, na, na
2, N, 0, 13.656, 4.19467e+12, 0
3, T, 0, 13.656, 4.19467e+12, 0
4, N, 0, 13.703, 4.19467e+12, 0
5, N, 0, 16.094, 4.19467e+12, 0
6, T, 0, 16.094, 4.19467e+12, 0
7, N, 0, 16.125, 4.19467e+12, 0
8, N, 0, 17.641, 4.19467e+12, 0
9, T, 0, 17.641, 4.19467e+12, 0
10, N, 0, 17.672, 4.19467e+12, 0
...

Best regards

Re: MIPTRACE (objective value not in the exponential format)

Posted: Tue Apr 23, 2019 9:32 am
by dirkse
nuno_r,

Reading your post, I had a few questions: what solver? what version of GAMS? Who is nuno_r?

I think your time would be better spent in scaling the model so the objective value is in a much smaller range (say, in the thousands instead of the trillions). The same goes for the constraints and variables. With the scale you have now, you cannot expect the solver to do its best work. Of course, you can always hope . . .

-Steve