Skip to main content

Hi,

I am working on a University Examination Timetabling Problem (UETP) using Binary Integer Programming, and I am running the model in AIMMS. The model is taking an unexpectedly long time to solve.

Previously, I set the optimality relative tolerance to 0.01, and while the results were not 100% optimal, I was able to obtain optimal solutions in a reasonable amount of time. However, when I set the optimality relative tolerance to 0, the model has been running for an extended period. After 5 to 9 days, the gap was 0.38%, and after 10 to 16 days, the gap only reduced to 0.37%, this model still running.

I ran this model on two different machines to compare performance, and both showed similar results:

  • Server: Intel(R) Xeon(R) Silver 4214 CPU @ 2.20 GHz (2 processors), 16 GB RAM, Windows 10 Pro
  • Desktop PC: 13th Gen Intel(R) Core(TM) i7-13700 @ 2.10 GHz, 64 GB RAM, Windows 11 Pro

My questions are:

  1. What could be causing such a long execution time with this model?
  2. Are the specifications of these machines sufficient to run this model efficiently, or could hardware limitations be a factor?

Thank you in advance for your assistance.

Hi @Wanie Saidi. The current model might run forever, if it takes a couple of days to go from a gap of 0.38% to 0.37%. (I guess the number of unexplored branch-and-bound nodes is still increasing.) A faster computer is not going to help you getting to 0%.

The cause of the long execution time is that this seems to be a challenging problem, and your model formulation might not be "optimal”. Therefore you might want to strengthen your model formulation, which is easier said than done (see here), or you might want to try a different model formulation, for example:

https://link.springer.com/article/10.1007/s10951-024-00805-0


Reply


Didn't find what you were looking for? Try searching on our documentation pages:

AIMMS Developer & PRO | AIMMS How-To | AIMMS SC Navigator