Skip to content

KeunJuSong/Plasmo_tutorial

Repository files navigation

Plasmo_tutorial

Foods for thought:

  • Hierarchical modeling? <= Use aggreagte method in Plasmo.jl to make coarse graph, and then first solve it fastly (in this step, let's check the number of variables and constraints). After that, using the solution of the coarse graph as the warm start to solve the "real" problem.
    • For doing that, we need to consider the decomposition options... (e.g., ADMM, Benders decomposition, Shcwarz decomposition, ... etc.)
  • Do we need to consider Multi-Stage or Multi-level optimization for solving large scale problem?
  • Optimization solvers that leveraging GPUs. => MadNLP
  • Which part that NN should involve to solve large scale problem?
    1. 이미 상용솔버에서 NR로 풀어서 전력망을 잘 운영하고 있으니 최대한 이걸 활용하는 솔루션
    2. NR 자체를 NN으로 대체하는 솔루션
    3. 분산/병렬최적화 (ADMM 등등)

About

Tutorial codes about Plasmo.jl

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published