Closed jokochems closed 1 year ago
@yannickwerner: Thanks you very much for reviewing https://github.com/pommes-public/pommesdata/pull/36 and https://github.com/pommes-public/pommesdata/pull/37! 😃
Would you please have a look at the adaptions here in pommesdispatch
as well?
Please, especially have a look at the changes in the subroutines for transformers minimum and maximum load. It seems that giving the model a difference between min
and max
of +0.01 is not enough and leads to an infeasible model. I therefore increased this to +0.05. Also note that there has been a bug in the prior version and the respective section for AT and FR hasn't worked as expected. I fixed this.
I also updated the tests which is why the changelist is so extensive. You can ignore that for now. I checked it: Tests are working and code coverage remains at 88%.
@yannickwerner: I added another very small update today since pommesdata
now includes a distinction between nominal and real cost values. We'll use the nominal ones.
@yannickwerner: Have you already had a chance to check it?
@yannickwerner This one should be finished now. I removed the workaround with the min/max margins again since we now address this properly in adjusting gradients. Please review again and merge in case everything is fine.
@yannickwerner Would you please review one (hopefully) final time? I suppose, in combination with the latest changes to pommesdata
, this one should now resolve the problems we faced with infeasibilities.
Closes #34 by introducing new costs values, time series for costs and availabilities