Jonas Hörsch
Jonas Hörsch
Closing, since it was already merged as part of #415
> This ambiguity you have right now as well. We just change the mapping. `n.generators.p_max_pu` -> `n.generators.static.p_max_pu` `n.generators_t.p_max_pu` -> `n.generators.varying.p_max_pu` Right now you could have in the static one just...
> Coming back to this point raised by @coroa > > > 1 Should Generators become the new full data container for what we have now at net.generators and net.generators_t?...
Note that with the implementation here, it is currently unclear how to remove an SOS constraint again or where to access dual values. Is that a thing for SOS constraints?
Indeed, i was only sketching out how it works (it's still a draft). Currently supported: - lp-polars - direct (for gurobipy) I don't think it's difficult to do the others.
For removing, it would be easy to add: ``` m.remove_sos_constraints(variable) ```
Open question: Should we use the `sos_dim` coordinates as weights of the SOS constraint or simply do not add weights? Context: The weight of an SOS constraint is only relevant...
2nd open question: should we add a check/raise when the solver does not support sos constraints in the lp file somewhere?
1. **weights**: Thanks for this quote of the gurobi docs, which confirms that the weights do not have an inherent meaning any more. For avoiding surprises and for ease of...
Rebased on latest master (ie. where only lp-polars remains) I added the representation that we talked about above. I added the `m.remove_sos_constraints(variable)`. Solver support seems to be: 1. `cbc` -...