Simplify adding new constraints

Digitalized-Energy-Systems / opfgym

A gymnasium-compatible framework to create reinforcement learning (RL) environment for solving the optimal power flow (OPF) problem. Contains five OPF benchmark environments for comparable research.

https://opf-gym.readthedocs.io

MIT License

2 stars 0 forks source link

Simplify adding new constraints #1

Closed thomaswolgast closed 3 weeks ago

thomaswolgast commented 1 month ago

Currently, adding new constraints is very complicated because some hidden methods need to be overwritten. Further, there is no simple way to remove the standard constraints (voltage band, line load, etc.).

Goal: The user should be able to easily define new constraints and add them to a given environment.

TODO:

Implementation of the new mechanism
Allow for removing existing constraints
Test for the new mechanism
Example that demonstrates the new mechanism on non-trivial constraints

thomaswolgast commented 1 month ago

Probably it's a good approach to define a Constraint or Violation object to achieve this, considering the three (!) return values currently in the base env class.

thomaswolgast commented 3 weeks ago

The default constraints should simply be extracted from the pandapower net