search

odow / SDDP.jl

A JuMP extension for Stochastic Dual Dynamic Programming
https://sddp.dev
Other
305 stars 61 forks source link

Does `SDDP.jl` support modeling with endogenous uncertainty and custom decision rules? #745

Closed WuSiren closed 4 months ago

WuSiren commented 4 months ago

First of all, I would like to extend my sincere admiration to the developers of this great package!

May I ask here some questions about the potential functionality of SDDP.jl?

Q1: Does the package support endogenous or decision-dependent uncertainty? If not currently supported, are there plans for future implementation?

Q2: Does the package allow for user-defined parameterized decision rules, such as an affine decision rule? If not, are there plans for future support? (What would be the complexity of adding such features? How long might it take to implement?)

Looking forward to your reply at your convenience! Thank you in advance!

odow commented 4 months ago

Q1: this directory mentions a not-yet public paper: https://github.com/odow/SDDP.jl/tree/master/papers/mdp. I'm not going to explain or answer questions on the code yet, but you're welcome to read it.

Q2: no support and no future plans. I don't have any idea about the complexity of adding such features, and I don't know how long it would take. The code is open-source though so you're welcome to make a fork and try something :smile:

odow commented 4 months ago

The issue at https://github.com/odow/SDDP.jl/issues/696 was more about simulating other decision-rules. We wouldn't provide a way to train LDRs, for example.

WuSiren commented 4 months ago

OK, I see. Thank you for your kind response and suggestions! 🤝🤝