Oscar Dowson

You can comment out the failing tests for now: https://github.com/ds4dm/Tulip.jl/pull/137

Here's what I came up with: ```Julia julia> import Enzyme julia> f(x...) = log(sum(exp.(x))) f (generic function with 1 method) julia> function ∇f!(g::AbstractVector{T}, x::Vararg{T,N}) where {T,N} g .= Enzyme.autodiff(Enzyme.Reverse, f,...

This should be closed. We're discussing it over in https://github.com/jump-dev/JuMP.jl/pull/3712.

Yeah, this is one part of the API that needs fixing. But I could never decide the right thing to do. The answer is that you need to use `Symbol(n)`...

One reason I haven't made this nicer is that, in general, I find that in most cases you want to use `SDDP.simulate` instead of querying particular values of the decision...

This will be easier to test and benchmark on some of the upcoming MSPFormat problems, in which SDDP.jl performs poorly compared with QUASAR. (A Markovian graph with 22,000+ nodes.)

One blocker is https://github.com/odow/SDDP.jl/blob/e631d22979163bc321908d97fdab53bcd7cbe26f/src/plugins/duality_handlers.jl#L64-L80. We'd need to change this so that we did it node-by-node, instead of all in one go.

I have something in #758 that seems to work well.

> ERROR: MathOptInterface.UnsupportedConstraint{MathOptInterface.ScalarNonlinearFunction, MathOptInterface.GreaterThan{Float64}}: `MathOptInterface.ScalarNonlinearFunction`-in-`MathOptInterface.GreaterThan{Float64}` constraint is not supported by the model. Gurobi does not support nonlinear constraints. Can you provide a reproducible example of your real problem? `1 /...

There is no way around this. You will need to do something like Andrew's paper, where you train a policy with a convex approximation of the inflow dynamics, and then...