Open
Description
I'm trying to get optimal performance for my setting where I have to generate about 25000 scenarios and solve these.
I tried almost all open source solvers including cbc, glpk, scip, highs, mbnb and also hexaly. I wrote my own template class to get a comparable behaviour. Hexaly is not as fast as expected but since I try to fit a MILP model to it, it could be expected. For other solvers, especially combined with pyomo, I set the temporary directory
for call of solvers to a RAMDISC
to gain a little bit of speed.
So I wonder, if there is any chance of setting the temporary path for python-mip when calling SCIP or HIGHS. It seems to be no such option available.
Metadata
Metadata
Assignees
Labels
No labels