Originally posted by: nr-mz

Dear Cplex,

I would like to write you to draw your attention on a question..

I am using Cplex 12.6.1. Actually, each time I run my model which is a mixed integer linear model,  I got different solution times. Is it so normal to have different solution time  for the same model with the same data in every run? Could you please help me on this question?

Thank you so much for your attention

Best regards

This is odd unless you set parameter CPX_PARAM_PARALLELMODE to -1. Did you?

If not, is CPLEX the only process running on your system or are other memory/CPU hungry processes running in parallel? Despite the different solution times, do you get the same solution value? Can you reproduce this by exporting the model to a SAV file and then repeatedly reading and solving it in the interactive optimizer?

Originally posted by: nr-mz

First of all thank you so much for your attention..

No, actually I dont use CPX_PARAM_PARALLELMODE in my code. Also, there is not another programming in running when I llaunch my code..

Yeah, despite the different solutions, I get the same values..

Actually, my work at the first step is to solve a project scheduling model in cplex and then (second step) to apply an improver algorithm.

For first step of job, I have created my opl project in cplex environment, ( That is to say  I use directly cplex interface and I do not implement cplex in java or other programming enivirements ). Then, when I solve my model in cplex, for every run I get different run times.. and it is quite strange for me.. Could you please guide me how i can teat with this issue? How interative optimizer will help?

P.S: For the second step of my work, i have another problem. I would like to ask within this post, please : https://www.ibm.com/developerworks/community/forums/html/topic?id=1ba72c86-3673-42e4-bf5c-86a51a3cd335&ps=25

Thank you so much for your attention

Regards

The interactive would help to narrow down where the time goes. Repeatedly solving the exact same file with the interactive optimizer should give similar solution times. If it does not then there is a bug somewhere in CPLEX. If it gives similar solution times then something odd is happening at the OPL layer or in the IDE.

Could you also post the log output of two runs that have different solution times? Maybe we can spot something.