We are tracking and investigating the benchmark instances and solvers that returned an error in this spreadsheet:
https://docs.google.com/spreadsheets/d/18bb8xqZhydMzGq7jXmAsmFk5PNzjwoTUMws5lHWHqJ4/edit?usp=sharing
The latest versions of each solver, used in the main ranking in the home dashboard, have 20 errors:
Most of these appear to be errors with the corresponding solvers, and we have opened issues to follow up with the solver developers.
Here are some other known errors/issues with previous solver versions (2023 and before):
We are tracking and investigating the benchmark instances and solvers that returned an error in this spreadsheet:
https://docs.google.com/spreadsheets/d/18bb8xqZhydMzGq7jXmAsmFk5PNzjwoTUMws5lHWHqJ4/edit?usp=sharing
The latest versions of each solver, used in the main ranking in the home dashboard, have 20 errors:
Most of these appear to be errors with the corresponding solvers, and we have opened issues to follow up with the solver developers.
Here are some other known errors/issues with previous solver versions (2023 and before):
unknowntermination condition when it terminates due to reaching the duality gap tolerance. This is a bug inlinopywhich has been fixed by us in the latest version, but the latest linopy is not compatible with old SCIP versions.warningstatus for some instances, and again we have fixed this in the latest linopy, but because of the above issue the 2023 conda env uses an older linopy version. When the above is fixed, this will also be fixed.unknowntermination condition when it reaches gap limit. Linopy needs to be fixed to report this correctly.