Oscar Dowson

This also appears in @michel2323's GO models. At the very least, we should be able to deal with objective functions that are separable, even ones with a subset of terms...

See also https://github.com/JuliaSmoothOptimizers/ExpressionTreeForge.jl cc @paraynaud

Currently, MathOptSymbolicAD exploits problems in which there is repeated structure in the Jacobian. So if you have a bunch of constraints `log(a[i] + x[i]) The experimental MadDiff does this well...

Perhaps to clarify, I don't intend to use ETF in this package. It was just a link I was chatting to @amontoison about that seemed relevant and adjacent. It has...

Here's the clnlbeam example for reference: https://github.com/SciML/SciMLBenchmarks.jl/pull/836

No. Once this LQOI is tagged, we should remove lines like https://github.com/JuliaOpt/LinQuadOptInterface.jl/blob/b553df3dde540981c702943b420e26716145715d/src/solver_interface.jl#L393 and remove the `objective_constant` field.

We add them as a block so it might be slightly more efficient: https://github.com/JuliaOpt/LinQuadOptInterface.jl/blob/ebd36b1a784330f21a0fec9ca4c1c3659210cefa/src/constraints/vectoraffine.jl#L28-L51 But the deleting etc happens row by row. @joaquimg might disagree, but I think the added...

The bridge should probably add them via `add_constraints`, then it is up to the solver to supply an efficient implementation. Then it is no problem. Direct mode is fine (if...

For `VectorAffine`->`ScalarAffine`, see https://github.com/JuliaOpt/MathOptInterface.jl/pull/413

> Would there be any benefit to adding a few basic solver traits as part of this? I am thinking of something that signifies columnwise(Clp) vs rowwise(gurobi) data structures. I...