Originally posted by: MeirLev

I have an Integer Programming model with ~1,300 variables and ~7,800 constraints, which I am running from the ILOG CPLEX Optimization Studio (version 12.6.2 under Ubuntu 14.04) with the Engine tab open. 

After 47 seconds, the following stats are shown:

Clique cuts applied:  14
Implied bound cuts applied:  518
Flow cuts applied:  176
Mixed integer rounding cuts applied:  178
Zero-half cuts applied:  25
Lift and project cuts applied:  5
Gomory fractional cuts applied:  2

Root node processing (before b&c):
  Real time             =    3.19 sec. (2433.05 ticks)
Parallel b&c, 4 threads:
  Real time             =   12.97 sec. (15760.99 ticks)
  Sync time (average)   =    1.11 sec.
  Wait time (average)   =    1.16 sec.
                          ------------
Total (root+branch&cut) =   16.17 sec. (18194.04 ticks)

Following this, the Studio's window becomes grey and it does not react to anything. Nothing new is displayed in the Engine window. After at least 15-20 minutes, the window wakes up and the solution is shown. 

What is the freezing behavior that I am observing and how can I get rid of this seemingly wasted time at the end?

 EDIT: When I run the same model from the command line using oplrun, the model is solved in 47 seconds...

Hi,

I guess some of the data in the problem browser is slow to compute.

Can you try to add a main block:

thisOplModel.generate();

cplex.solve();

thisOplModel.postProcess();

regards

Originally posted by: MeirLev

I added this:

main {

    thisOplModel.generate();

    cplex.solve();

    thisOplModel.postProcess();

}

Still, after the problem was solved, the freeze followed...

Hi,

and then if you comment 

thisOplModel.postProcess();

?

You may also try 12.6.3

You may also attach your model and data so that other forum users have a look.

regards

PS:

do you have many dexpr in your model ?

regards

https://www.linkedin.com/pulse/low-barrier-entry-optimization-through-cplex-alex-fleischer

Originally posted by: MeirLev

Commenting out thisOplModel.postProcess(); did not make a difference.

There are 129 dexpr (one of them is being maximized), each of them being a sum of 64 variables...

Hi,

can you try the workaround given at

http://www-01.ibm.com/support/docview.wss?uid=swg1RS00856

?

Do you have a postProcess block ?

Can you post your model and data here ?

regards

Originally posted by: MeirLev

Sorry, a typo: only one of the expressions has as many as 64 variables in the summation, the others are sums of 6 variables.

I don't believe memory can be a problem (though I know that the Studio uses only 32-bit addresses, 129 expressions can't occupy 4 GB, can they?) Also, the only thing my execute block currently does is printing the value of the objective function.

Unfortunately I can't post the model, since it constitutes original research.

Hi,

so you could log a PMR to get some support.

Maybe you could use a 64 bits version.

What happens if you remove the postProcess part ?

regards

Originally posted by: MeirLev

Log a PMR? Do you mean simply contacting support and describing my problem or creating a log of what happens in the Studio? If it's the latter, could you please provide a link to where this feature is explained?

Yes, oplrun (which uses 64-bit addresses and does not rely on Java) does the job. My production runs will use it. However, it is may be convenient to use the IDE for the development stage.

No change in behavior when the execute block is removed -- long freeze after the problem has apparently been solved.

Hi,

then my suggestion is to log a PMR at https://www-947.ibm.com/support/entry/portal/support

regards

Originally posted by: MeirLev

OK. Thank you very much for your help.