Open ckaldemeyer opened 1 year ago
Here you can find a performance benchmark compared to Gurobi 8, CBC and GLPK with SCIP version 7 (slide 18): https://www.gurobi.com/wp-content/uploads/Gurobi_8_Performance_Benchmarks.pdf?x58432
And here some comparisons to different other solvers on test suites: https://mattmilten.github.io/mittelmann-plots/
I did not test it, yet, as I did not know that it has a usable license now. Do you think it's possible to have an hands-on at the next user meeting? (Currently, I doubt that my callendar allows to do this without tutoring or an explicit explanation.)
I won't be on the next user meeting as I am on parental leave at the moment. Wouldn't it be sufficient to test the current examples and optionally compare the results?
I have tested some of the MIP examples successfully and created a PR in order to discuss, contribute the documentation on how to install it and check how it can be tested.
https://github.com/oemof/oemof-solph/issues/855 is closely connected and might be integrated in the same PR -> see my PR
Hi everybody,
I just wanted to mention that the SCIP solver is licensed under the Apache 2 license from version
8.0.3
onwards and can be used to solveLP/MIP
problems.My experiences are so far very good in terms of usability and performance when compared to
GLPK
andCBC
and in the newer versions, it can be used easily from Pyomo using AMPL bindungs, which I could successfully test under Linux and Windows 10. See my related question here: http://listserv.zib.de/pipermail/scip/2023-March/004601.htmlIf you think, this provides some value and your tests are positive too, I would offer to add some documentation on how to install it on Linux und Windows.
Best wishes, Cord