Matlab Error "model.A is not representable"
AnsweredDear Gurobi community,
I'm using Gurobi's implementation of ILP in Matlab and when calling the solver, I get this error message in Matlab:
"Error using gurobi
model.A not representable
Error in intlinprogGurobi (line 167)
result = gurobi(model, params);"
I've tried to find an explanation and I couldn't find any on google or in the Gurobi documentation. Any hint to help me to know what's going on ?
Thank you in advance !
-
Official comment
This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try our AI Gurobot?. -
Hi Julien,
Could you share a sort, self-contained code snippet generating the error you are getting? Also, could you share the size of your constraint matrix (number of rows, number of columns, number of non-zero entries)?
Thanks,
Elisabeth
0 -
Hi Elisabeth,
This is based on data on a secured server so that it is hard for me to provide any reproducible code.
The problem is big I guess (the server can handle the storage and has calculation power):
220,938,496 variables
30,314 inequality constraints (450,587,296 non zero entries, coded as sparse matrix)
1,661 equality constraints (8,810,438,174 non zero entries, coded as a sparse matrix)
I understand that this might be too big/long to solve but, you know, we never know for NP Hard problems :) The same code with a smaller dataset runs in about 2h so I was surprised to see a crash with the bigger dataset. Especially that I thought that, in the worse case scenario, it might just never stops but I was not expecting an error code.
So I wanted to check whether I can still hope to make it run or whether this error somehow means "too big just forget about it" but since I couldn't find any documentation on such specific error message, I was puzzled.
Thank you for taking the time to answer.
Best,
Julien
0 -
Hi Julien,
The error comes indeed from the huge problem size. For the Matlab interface, there is currently a limit of 2 billion non-zero entries in the constraint matrix, which is largely exceeded by your model. Increasing this limit is currently registered as a feature request.
Best regards,
Elisabeth
0 -
Thank you very much for your answer! Good to know that I can move on now :)
0
Post is closed for comments.
Comments
5 comments