Benoît Legat

Results 487 comments of Benoît Legat

> We'll loose the ability to rewrite the boolean expressions via bridges, but we might be able to write a layer that transform a ScalarNonlinearFunction that represents a boolean expression...

We could have `@constraint(model, )` create a `ScalarNonlinearFunction`-in-`MOI.EqualTo{true}` and in the bridge from `ScalarNonlinearFunction`-in`MOI.EqualTo{Bool}`, we can check that the nonlinear function is a boolean expression. Otherwise, we can throw the...

Yes, for vector set we could wait to see if there is a real need for it. But we could at least support it for abitrary scalar set. https://github.com/jump-dev/JuMP.jl/pull/3184/ is...

> If any future reader has a strong wish for the vector syntax `@variable(model, x[1:3, 1:4] in SecondOrderCone())` please comment and we can re-open this issue. For the record, @araujoms...

- [x] #3250 quantum information, state discrimination suggested by @araujoms in https://discourse.julialang.org/t/error-when-constraining-variable-to-be-hermitian-psd/95223

- [x] #3259 A tutorial playing with adding bridges, removing bridges and printing active bridges. Probably an examples with RootDet -> GeoMean -> (3 possible bridges for GeoMean) -> SCS...

> This question comes up surprisingly often: > > https://discourse.julialang.org/t/connecting-a-simple-first-order-solver-to-solve-standard-form-linear-program-to-jump/95694 > > I can see it being a good feature in `Utilities`, or as a tutorial with Simplex? It'd also...

Tutorial on using `Dualization.dual_optimizer` (e.g., close to https://jump.dev/JuMP.jl/dev/tutorials/conic/introduction/#How-to-choose-a-solver) and update conic tutorials to make sure they don't use pure standard conic form with geometric conic form solver or vice versa....

I was a bit worried about the argument raised yesterday that it won't work with Array. I think most users don't even notice the change of Containers at the moment

These might indeed be related but I think it's different. I guess `operate!` should mutate the coefficient if the terms have the same monomial and error otherwise. This simply need...