search

floracharbo / MARL_local_electricity

Multi-agent reinforcement learning for privacy-preserving, scalable residential energy flexibility coordination
GNU Affero General Public License v3.0
23 stars 5 forks source link

Clean up network optimization #54

Open julie-vienne opened 1 year ago

julie-vienne commented 1 year ago

@floracharbo now that we have several options to solve the optimization (subset of lines or iteration) and these can be supplemented with compare_pandapower_optimisation, I think it could be useful to clean up the different options and make them more comprehensible. This includes printing meaningful summary statistics (so far, a summary is printed and a txt file can be created) and decide on which methods and combinations make sense, e.g. no need to do an iteration and compare_pandapower_optimisation = True since it's already included.

julie-vienne commented 1 year ago

I believe this has been tackled in PR #57. In the iterative approach, the last iteration results is anyways corrected with a pandapower load flow and in the subset approximation method, there are post-processing tests in place to correct optimization results if needed. If compare_pandapower_optimisation = True, then these corrections are performed anyways and allow a comparison of the subset optimization and the load flow for each time step. @floracharbo unless you think this still needs improvements, I will close this issue as completed :)