Question about MibS Solver (potential Bug)

[NKonti]

Julia_Model_MibS.txt

Hello, I am a student at the University of Siegen. I am using the MibS solver for my Master's thesis. However, I trouble with the Julia model in the appendix. In particular, the output of the lower level variables does not reflect what the lower constraints should do. Actually, each node should be visited once and no circles to itself should be created. Is this a bug?

Best regards,

Nick Kontorzik

[tkralphs]

It would be helpful if you could add some details, such as

• a plain text version of the model,
• the output from running the optimization,
• what the optimal solution reported by MibS is, and
• what you think the optimal solution should be.

Then we can at least see whether this is indeed a bug or perhaps a misunderstanding on your part.

If it does seem to be a bug, then it would be helpful if you could put the model into the native format of MibS, which is an MPS or LP file plus an auxiliary file specifying which variables and constraint are lower level (see https://coin-or.github.io/MibS/input.html) so that we can try to reproduce directly with MibS and eliminate the possibility of a bug in the wrapper.

[NKonti]

Thank you for your quick reply. Ok, apparently you can't access my txt file attached to the previous message. I hope you can now open the following file with the model and the necessary information.

Model_MibS.txt

[tkralphs]

I could read the text file, I was just hoping you could create both a human-readable version of the model (like you would write in a paper) and the input files that would allow me to solve directly in MibS. This would allow me to quickly try a more recent version and see if there is a bug that was fixed. The version of MibS that is installed with MibS_jll is still an older one. Hopefully, t will be updated soon.

I could of course do this myself, but I'll get to looking at this a lot faster if you can help me out with this additional information.