0

enter image description here

I am running a code with millions of decision variables and close to 60000 constraints. I added 7000 new constraints and am receiving this error - CPLEX Overflow occurred.

I doubled the memory of the virtual machine where I am performing the computation, but am still facing the same problem. What could be the issue? I am still getting an output. But I do not know if this is the optimal output or not, because the profiler tab shows the error message "Overflow occurred. Use oplrr.

Sanchita
  • 1
  • Hi,if you get an output then maybe you re fine and if you really need the result of the profiler then use oplrun -profile (in command line) – Alex Fleischer Dec 21 '19 at 15:04
  • Thanks for your response Alex. I am doubtful, this is the correct answer because the code takes 4.5 to 6 hours to run. On adding additional constraints (close to 7000), I am getting an answer within 45 minutes, with a profiler message "Overflow occurred" which is abrupt. When run on a virtual machine with larger memory, the code runs for 2 hours and then shows the message. Also how to use opl run profile (in command line)? Sorry, if this is a basic question. New to this. – Sanchita Dec 22 '19 at 17:01

1 Answers1

0

Suppose you model is mod.mod and your data is dat.dat then in the command line you could write

oplrun -profile mod.mod dat.dat > prof.txt

and then the profile will go to prof.txt

See this page for oplrun documentation. And your issue is addressed in this technote.

Problem

When running model through OPL IDE, the profiler gives the output "Overflow occurred, please use oplrun -profile"

Resolving The Problem

There is a limitation in what the OPL IDE can display because of 32bits Java integer specification because of which the profiler will not be able to display the results. In such cases, to obtain the profiler ouput, use oplrun with the "-profile" option.

Community
  • 1
  • 1
Alex Fleischer
  • 9,276
  • 2
  • 12
  • 15