Oscar Dowson
Oscar Dowson
> (e.g., being real) But this was the entire point. We have no way of knowing if an arbitrary `MOI.AbstractFunction` satisfies the required properties, or the required `promote_op` etc. At...
Kinda agree. I see both arguments for MOI 2.0. But for now, disallowing bridges unless functions implement `is_maybe_complex` is a breaking change, so it needs to be `is_maybe_real`.
I'm also okay to make this function private for now.
Is there anything left to do here?
I confirmed that this is a problem in the C API, not the Julia package: https://github.com/jump-dev/Cbc.jl/issues/211#issuecomment-1347050413 But @SBuercklin has a bigger problem: their model is poorly scaled, which I guess...
Is there anything actionable to do here? I don't think we should compile with a different value, and the overrides work.
There is https://github.com/cddlib/cddlib/issues/34, but it doesn't explicitly ask for it to be a parameter
💯 for this. > The last one being less trivial, two options would be: Vote for Complementarity constraints. I talked to M. Ferris about this and we're super keen to...
Smaller example is ```Julia julia> using JuMP, Ipopt julia> import ParametricOptInterface as POI julia> function main() model = Model(() -> POI.Optimizer(Ipopt.Optimizer())) @variable(model, z in Parameter(0)) @variable(model, x) @NLobjective(model, Min, x)...
Perhaps `POI` should error if it detects a `NLPBlock`.