Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
Hello, sorry for disturbing.
In the model i am runing for my master thesis, there is 2 scenarios where i am having the same error related to time limit and i do not really know why since this scenarios just have a small change regarding the other scenarios.
here is the error:
I am also uploading the lst file and runlog of this scenario, please if someone know show can this be solved i would really appreciate an answer.
master_thesis_ppa.zip (Size: 78.2 KB / Downloads: 4)
Posts: 54
Threads: 4
Likes Received: 0 in 0 posts
Likes Given: 3
Joined: Apr 2022
Hello,
it seems that you reached the time limit, you can modify the value RESLIM in the run manager from 50000 to a bigger value.
Another solution is to deactivate the crossover which takes most of the time, barrier algorithm takes around 800s in your case. You will need to modify your cplex.opt.
Based on the computer you have you can change the Threads=7 to Threads=-1 in cplex.opt to use 15 out of the 16 threads you have. It will speed the barrier resolution time.
this documentation may be interesting for you:
https://iea-etsap.org/webinar/CPLEX%20op...models.pdf
Victor
Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
(19-04-2024, 07:32 PM)VictorG Wrote: Hello,
it seems that you reached the time limit, you can modify the value RESLIM in the run manager from 50000 to a bigger value.
Another solution is to deactivate the crossover which takes most of the time, barrier algorithm takes around 800s in your case. You will need to modify your cplex.opt.
Based on the computer you have you can change the Threads=7 to Threads=-1 in cplex.opt to use 15 out of the 16 threads you have. It will speed the barrier resolution time.
this documentation may be interesting for you: https://iea-etsap.org/webinar/CPLEX%20op...models.pdf
Victor
Hello Victor, I do not understand the cplex.opt thing i should change. Where can i find that document? because in the pdf you have send me it does not explain what exactly to change and where
Posts: 54
Threads: 4
Likes Received: 0 in 0 posts
Likes Given: 3
Joined: Apr 2022
Have you tried to increase RESLIM ? does it work ?
cplex.opt is related to solver options in the run manager:
https://veda-documentation.readthedocs.i...definition
Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
(22-04-2024, 05:58 PM)VictorG Wrote: Have you tried to increase RESLIM ? does it work ?
cplex.opt is related to solver options in the run manager: https://veda-documentation.readthedocs.i...definition
No i have not tried. But how can i change that thing in the cplex.opt? because that just appear in the run manager. But how can i change the threads to -1??
Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
(22-04-2024, 05:58 PM)VictorG Wrote: Have you tried to increase RESLIM ? does it work ?
cplex.opt is related to solver options in the run manager: https://veda-documentation.readthedocs.i...definition
The thing is that i do not find where to change those threads from 7 to -1
Posts: 54
Threads: 4
Likes Received: 0 in 0 posts
Likes Given: 3
Joined: Apr 2022
You can find opt files in 'Veda\Veda_models\my_model\AppData\SolverOptFiles' with the path adapted to your context.
Modifying the threads number will mostly speed up barrier phase but not the crossover so it won't probably solve your RESLIM problem
Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
(23-04-2024, 12:24 PM)VictorG Wrote: You can find opt files in 'Veda\Veda_models\my_model\AppData\SolverOptFiles' with the path adapted to your context.
Modifying the threads number will mostly speed up barrier phase but not the crossover so it won't probably solve your RESLIM problem
Thank you so much. Is strange because in my model i do not have CPLEX in the solver options:
Posts: 54
Threads: 4
Likes Received: 0 in 0 posts
Likes Given: 3
Joined: Apr 2022
Posts: 59
Threads: 12
Likes Received: 0 in 0 posts
Likes Given: 0
Joined: Feb 2024
(23-04-2024, 01:03 PM)VictorG Wrote: just modify BA121-7 then
Hello Victor,
Sorry for disturbing you again. But I increased my RESLIM to 90000 from 50000 and it still is not able to solve it. I have no idea why this may be happening, specially since this is just one scenario from 5 I have run and the others finished relatively fast. Do you find anything else I could do? Thank you so much
master_thesis_ppa.zip (Size: 96.28 KB / Downloads: 2)
Posts: 54
Threads: 4
Likes Received: 0 in 0 posts
Likes Given: 3
Joined: Apr 2022
24-04-2024, 12:28 PM
(This post was last modified: 24-04-2024, 12:29 PM by VictorG.)
It is difficult to help without having a look at your model.
Based on your description, if you have a parent case that solves quickly then what you added in your variant scenario is probably responsible for this time solution increase. You should have a look on your modification.
Sometimes dummy imports are responsible of numerical issues. What you can do is go in veda in Tools>User options>Syncing options and you uncheck the four boxes. Then start from scratch, sync your model and run a case. If there is an infeasibility, the log will give you information of the elements of your model involved, based on this you can alter your model to remove it.
It is a bit of trial and error