Originally posted by: MachexC

Hi Alex,

My model run with very long time, I was thinking of some ways, spending a month but could not solve. Please attached an engine log, the gap of 100%. Thus, could you give me some advice? 

Thank you for help.

Mac

Hi,

are you a commercial user ?

Is your company in Colombia ?

regards

Originally posted by: MachexC

Hi Alex,

At the present, I am using CPLEX, academic from IBM.

No. I come from Vietnam,  Colombia is from my friend.

Please attached an engine log, the problem is that the "Best Integer" converges and "gap" very slowly, so it can't finish.
could you help me? I am really thank you.

Bests and regards,

Mac

Originally posted by: MachexC

Hi Alex,

Please help for this problem.

Thank you,

Bests,

Mac

Hi,

have you tried what @DanielJunglas 548c249b-bb6b-42f6-99b8-f0376c80829e suggested at https://www.ibm.com/developerworks/community/forums/html/topic?id=9ecaf017-7b51-4861-9ebd-3cb8fb90714c&ps=25

?

regards

Originally posted by: MachexC

Hi Alex,

Because the "Best Integer" is very big with 1e-11, so I made scaling the objective by 1e-11 to get results. But when I compare the results of the original model with and scaling model, in this case, the data is small size. I see that the results are different, the objective function of the original model is better (smaller than). 

@DanielJunglas said, "note that scaling the objective affects the absolute gap termination criterion!".  It is true, but I do not know how to solve it?

Thank you,

Bests,

Mac