search

jump-dev / ECOS.jl

A Julia interface to the ECOS conic optimization solver
https://github.com/embotech/ecos
Other
41 stars 17 forks source link

support for ecos_bb #38

Closed mlubin closed 3 years ago

mlubin commented 8 years ago

Closes #16 Tests will be failing until a new MPB is tagged (for conicSOCINTtest()) and https://github.com/embotech/ecos/issues/141 is fixed.

matbesancon commented 3 years ago

I assume this should be restarted with MOI, since the PR is a tad old?

matbesancon commented 3 years ago

It seems https://github.com/embotech/ecos/issues/141 hasn't been fixed, could this be solved with ecos_bb included in ECOS_jll? I'm not sure how it should interact with the main ecos shared library

mlubin commented 3 years ago

I assume this should be restarted with MOI, since the PR is a tad old?

I don't think it should be restarted at all. ecos_bb was never a competitive solver, and CVXPY has stopped using it by default because of correctness issues.

CVXPY comes with ECOS_BB – an open source mixed-integer nonlinear solver – by default. However ECOS_BB will not be called automatically; you must explicitly call prob.solve(solver='ECOS_BB') if you want to use it (changed in CVXPY 1.1.6). This policy stems from the fact that there are recurring correctness issues with ECOS_BB. If you rely on this solver for some application then you need to be aware of the increased risks that come with using it.

https://www.cvxpy.org/tutorial/advanced/index.html#mixed-integer-programs