DSDA_Improvement #11
DSDA_Improvement #11
Conversation
| @@ -152,7 +152,7 @@ def problem_logic_batch(m): | |||
| model_args={}, | |||
| starting_point=starting_point, | |||
| ext_dict=ext_ref, | |||
| mip_transformation=True, | |||
| mip_transformation=False, | |||
There was a problem hiding this comment.
I see that you removed the MIP transformation for each one of these steps. I would like you to check each subproblem, what the instance looks like, and what that MIP transformation is introducing. Start with small_batch.
| @@ -163,7 +163,7 @@ def problem_logic_batch(m): | |||
| iter_timelimit=timelimit, | |||
| timelimit=timelimit, | |||
| gams_output=False, | |||
| tee=globaltee, | |||
| tee=False, | |||
gdp/dsda/dsda_functions.py
Outdated
| @@ -11,7 +11,7 @@ | |||
| from gdp.dsda.model_serializer import StoreSpec, from_json, to_json | |||
| from pyomo.common.errors import InfeasibleConstraintException | |||
| from pyomo.contrib.fbbt.fbbt import fbbt | |||
| from pyomo.contrib.gdpopt.data_class import MasterProblemResult | |||
| # from pyomo.contrib.gdpopt.data_class import MasterProblemResult | |||
There was a problem hiding this comment.
If this is not imported anywhere, do not just comment it out but remove it
There was a problem hiding this comment.
Okay, I removed the imports
|
I made a separate environment containing Pyomo with Zedong's version of the LD-SDA code. I added the benchmark code and its results. |
|
These do not correspond to the results in the paper and there is no clear comparison between this implementation and the one from Zedong. My recommendation is to keep this repository as close to the paper as possible. |
I commentted out the dsda_function.py import MasterProblemResults. The current pyomo.GDPopt does not have data_class.py anymore. Moreover, I set the mip_transformation as False. There are issues on the external_ref() function in dsda_function.py