Oscar Dowson

@matbesancon see this error: https://discourse.julialang.org/t/how-to-implment-logsumexp-function-in-jump/84376/6

Oops. I always get this wrong. The Mosek and MOI conventions for Exponential cone are flipped. Try: `[y[i, j]*x[i] - t, 1, u[i]] in MOI.ExponentialCone()`

It's saying the problem is unbounded. Are you sure you don't mean ``` @variable(model, u[1:N]) @variable(model, t) @constraint(model, sum(u)

Ah. I fixed this one recently: https://github.com/jump-dev/MathOptInterface.jl/pull/1941 As a work-around, just add this method to your code: ```julia MOI.Utilities.set_with_dimension(::Type{MOI.ExponentialCone}, dim) = MOI.ExponentialCone() ```

This is where I'm not sure, sorry. I've never used DiffOpt, or dug into how it works. My guess is that it's still missing some features for ExponentialCone properly, or...

He means like this: https://github.com/jump-dev/MathOptInterface.jl/blob/master/test/functions.jl The nested testset approach of Julia is pretty sub-optimal. These functions are much nicer to write, are more self-contained, and we can easily extend to...

> You mention in the comments it is a realization of uncertainty that will be defined below It's the same ω as this one: ```julia SDDP.parameterize(subproblem, Ω) do ω y...

Renamed this to reflect that we should add documentation on different RHS processes. I'm unlikely to do this, so if someone reading has an AR example and wants to add...

@bernardokp would also like a GARCH one where the variance is accounted for: https://github.com/odow/SDDP.jl/issues/504

There is this, https://odow.github.io/SDDP.jl/stable/guides/improve_computational_performance/#Initialization-hooks, but @adow031 wants to integrate it into JuDGE.