CPLEX: Optimal solution with unscaled infeasibilities is thrown out

[thomgeo]

Hi everyone, I have an issue with CPLEX. for some problems, which Highs can solve, I get an unknown status from linopy, with the message that the problem is optimal with unscaled infeasibilities. As I understand, CPLEX introduced the solution of the scaled problem to the unscaled problem, and there were some slight infeasibilities, which usually do not mean a lot (see for examle here: http://iea-etsap.org/forum/showthread.php?tid=162)

Linopy, however, seems to not import this solution, potentially because the status is unknown and not optimal. Would it be possible to change this, so that linopy simply imports this solution, even though there might be slight infeasibilities?

Best,
Georg

[thomgeo]

Hi Fabian,

Thanks for the quick follow up. I updated linopy now, but it still does not import the solution if the status is "optimal with unscaled infeasibilities". Seems like it is still treated as infeasible.

Best,
GEorg

[tburandt]

Hi @FabianHofmann , I get the same problem with linopy 4.2.0. Have you found time to look into this again?