Veda2.0 Released!


Unknown error
#1
Dear Mr/Mrs,

I have a problem running a TIMES model. The model was succesfully synchronized and no errors were shown. After solving it using CPLEX engine I found out that the data acquired are not viable. Then I tried to add another parameter inside a process defined in the model (green mark inside the attached Excel file) and running the solver afterwards. This time an error occured meaning that the new parameter made the model infeasible.

Could you please give me any advice on that? It appears to be some kind of general problem of the whole model instead of singular process definitons or so.

There are two attachments showing the Excel table with the parameter newly added and the error logs.

Thank you.

Kind regards,
Lukas Novak


Attached Files Thumbnail(s)
       
Reply
#2
Ok, good that you have set "names 1" in your Cplex.opt file, to get much better readable info on the infeasibility.

My quick guess is that your problem may be with the ACT_BND staying constant while the capacity is decreasing and so cannot provide that much activity.  It would be safer to use AFA~LO instead.

I cannot tell for sure what the problem is, because you are not showing all the parameters for the process in Browse, you are not showing the period definition etc. etc.
Reply
#3
Dear Antti,
thank you for your advice. The model really could not provide that much activity using the given values of ACT_BND.

But the general problem with the model still remains. After succesfully solving the model, more than half of the data is missing. To be exact, some of the values are missing or they are zeros and there is less than half of used attributes shown in the results. It seems to me that there is some kind of general problem in the settings. Could you please take a look at the sheets of SystemSettings and BaseTrans (screenshots in the attachment) and tell me if they are fine with VEDA 2.0?

I also came across this one during the solving process: "LP Presolve eliminated 206089 rows and 281419 columns.". Couldn´t this mean for an instance that this amount of data had to be taken out of the model because they are wrongfully interconnected? It would explain a lot considering the results...

Best regards
Lukas Novak


Attached Files Thumbnail(s)
                           
Reply
#4
I am not a VEDA expert, and so I will give my answer only about the following:

> I also came across this one during the solving process: "LP Presolve eliminated 206089 rows and 281419 columns.". Couldn´t this mean for an instance that this amount of data had to be taken out of the model because they are wrongfully interconnected?

No, this is just an informative message from Cplex, saying how well it performed in its presolve routine, while reformulating the problem. With the best solvers, the solver presolve routines can often eliminate even most of the constraints and/or variables in LP models, while keeping them mathematically equivalent. In your case, the solver reported that it was able to eliminate a few hundred thousand constraints and variables. There is no reason at all to worry about it, unless you feel the numbers are far too small. The elimination does not take out any input data.

I hope the VEDA experts can help you with the VEDA2 issues.
Reply
#5
Alright, Antti, thank you very much. Then it really appears to be some kind of issue with the VEDA 2.0. Hopefully the experts would recognize the problem.

Best wishes
Lukas Novak
Reply
#6
Dear all,

I am still struggling to obtain proper results from the model. I am trying to find out if there is some kind of general problem with the model settings or definitions of various processes are wrong. I attached screenshots of parameters concerning one of the processes. Could you please tell me if there is some kind of mistake? I also attached shot of dummy imports, they might bring some further information.

Best regards
Lukas Novak


Attached Files Thumbnail(s)
           
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)