が出来たそうです。
Highsは、IvetGalabovaさんが去りMIPソルバの開発にブレーキがかかったそうですが、別に内点法ソルバのFilippo Zanettiさんが加わったそうです。
Highsは、行制約の殆どない多種重みの場合、例えば、kindle本/講師単価.nurse3ではAlgorithm1を凌駕する、SCIP並みの優れたソルバです。これで、MIP solution callback · Issue #896 · ERGO-Code/HiGHS · GitHubがサポートされたらよいのですが。
以前、教授に内点法ソルバのマルチスレッド化をお願いしたのですが、あんまり効果ないよ、と言われて自前で実装しました。進展があってReplaceできるといいですね。
highspyがどんなものか分かりませんが初夏には、ドキュメント化するそうです。これでまたファンも増えると思います。
*News*
The most recent news is that we have finally started to create something
approaching formal documentation <https://ergo-code.github.io/HiGHS/> for
HiGHS. It's still a work in progress, but it gives some information that
users have wanted for years, such as documentation of options. The detailed
documentation will relate to highspy <https://pypi.org/project/highspy/>.
This is a very cool, thin wrapper around the C++ Highs class that allows
HiGHS to be called from Python. Sorry, but those calling HiGHS from C or
C++ will have to make do with header files or documentation generated
elsewhere.
For the past six months the active HiGHS team has been just me and Ivet
Galabova. However, we have recently brought in Filippo Zanetti, as we look
to develop an improved interior point solver.
We dropped to just two people when our MIP developer, Leona Gottwald, left
at the end of September. This has put quite a brake on MIP support and
development, but we have plans to catch up with both.
HiGHS v2.0, with a full highspy interface and completed documentation
should appear by the early summer which, in Scotland, could be any time!
Best wishes