Originally posted by: SystemAdmin

Hello,

I use CPLEX to optimize several models from Java. It works fine in all cases except one when CPLEX just seems to "freeze" just before it is finished. This specific model which works fine otherwise, in one case stops at the line

6600 1277 1926.2198 30 1927.0000 1926.0886 37680 0.05%

after less than a second, a tree size of 0.3 MB, and 7 solutions found. If I test this model and data in OPL studio it is optimized within seconds. Anyone ever experienced strange behavior like that? I use CPLEX 12.1.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: EdKlotz

> CarstenBöhle wrote:
> Hello,
>
> I use CPLEX to optimize several models from Java. It works fine in all cases except one when CPLEX just seems to "freeze" just before it is finished. This specific model which works fine otherwise, in one case stops at the line
>
> 6600 1277 1926.2198 30 1927.0000 1926.0886 37680 0.05%
>
> after less than a second, a tree size of 0.3 MB, and 7 solutions found. If I test this model and data in OPL studio it is optimized within seconds. Anyone ever experienced strange behavior like that? I use CPLEX 12.1.
What happens if you call IloCplex.ImportModel just before the relevant solve() call that results in the freeze? If you export the SAV file of the problematic model and solve it with interactive CPLEX, do you get the same results as with OPL?.

Also, try setting the MIP interval parameter to 1 for the run in question (from Java, this means calling cplex.setParam (IloCplex.IntParam.MIPInterval, 1)) to see the point at which the OPL (or interactive CPLEX) run starts to diverge from the run in your program.

Also, try running the MIPex2.java program in your SAV file. That should help further isolate where the source of the trouble is (i.e. is it specific to Java, or specific to your Java program).
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

Hello Ed,

thanks for your answer. I haven't been able to test your suggestions, but I'll post an update in case I find out anything that might be useful to others. So far I know that to my surprise a set time limit is still adhered to. It just seems strange that it takes a couple of seconds to reduce the gap to 5% and then there are several minutes without any feedback.
#CPLEXOptimizers
#DecisionOptimization