Benoît Legat
Benoît Legat
Not at all, please do :) IIRC for the best performance, the double description method shouldn't restart the computation from scratch when intersecting with an halfspace. For this reason, [bensolve](http://bensolve.org/)...
I like the idea, let me try to extend it a little bit. Maybe we could do something similar to what `git commit` does which is, - when no title...
So ``` $ ... -t "Title" -b "Body" ``` would have the title "Title" and the body "Body" ``` $ ... ``` would launch a text editor. The first line...
@odow Convex transmit the problem as a SOC constraint and we don't have any bridge to transform this SOC constraint back to a quadratic objective as OSQP requires (and it...
What is the problem you have for BigFloat? MultiFloats don't implement MutableArithmetics.jl
An example of implementation can be found in mosek source code in `tools/examples/fusion/python/lownerjohn_ellipsoid.py` with functions `geometric_mean` and `det_rootn`.
The fix would be to encode it as an MOI [`GeometricMeanCone`](https://www.juliaopt.org/MathOptInterface.jl/stable/apireference/#MathOptInterface.GeometricMeanCone) constraint directly and not do the transformation in Convex.jl. Then the [`GeoMeanBridge`](https://www.juliaopt.org/MathOptInterface.jl/stable/apireference/#MathOptInterface.Bridges.Constraint.GeoMeanBridge) would transform it into SOC constraints if...
> Who would be negatively impacted by having the 0/1 bounds written as linear constraints? This is destroying structure of a model without knowing whether the underlying model handles variable...
The breaking part is not important so I won't argue with that. The reformulation is destroying structure, that is my issue with it. Bridges only destroy structure when it is...
I don't know the internal of all solvers at the moment (even less in the future). Maybe Juniper ? It's a Branch & Bound solver so it should be treating...