search

SECQUOIA / gdplib

An open library of Generalized Disjunctive Programming (GDP) models
BSD 3-Clause "New" or "Revised" License
9 stars 7 forks source link

Fixing typos and adding documentation to hda file #26

Closed parkyr closed 5 months ago

parkyr commented 6 months ago

first round of documentation to hda file, including header and function documentation. some typos were fixed as well in this pull request.

bernalde commented 6 months ago

Because of the recent PR merged by @ZedongPeng there are conflicts in this branch now. I would suggest you to figure it out with @parkyr before getting this through review

ZedongPeng commented 6 months ago

No problem. I will help Yirang resolve the conflicts.

ZedongPeng commented 6 months ago

Great. You made it. @parkyr

parkyr commented 5 months ago

I ran the code after the changes. The code runs, but it's giving me a slightly different obj value from the slidedeck; objective value for slidedeck: 5966.51; the objective value result: 5801.27.

I compared it with the the original (unmodified) file from master branch, and the values match. The code was not affected from the documentation modifications.

bernalde commented 5 months ago

This might arise from the subsolver that you are using, and whether it has improved since the time that previous table was made. Together with @ZedongPeng try to run all the different solvers to replicate the same results as in the table

ZedongPeng commented 5 months ago

Hi @parkyr . Have you run the code before the changes? Does the optimal solution (before changes) equal 5966.51 or 5801.27? You can go back to the old code with git checkout master.

parkyr commented 5 months ago

Hi @ZedongPeng, yes I did run the code from the master branch and it gives me 5801.27. I'll do some digging and see if I can recreate while you are away.

ZedongPeng commented 5 months ago

Hi @parkyr . Have you resolved this problem? I diff-checked the code line by line and didn't find any change that would result in a different result. Have you compared the log of GDPOpt before and after the changes? If you upload the logs in this PR, I can help you check them.

Btw, if you are using two different Python virtual environments in Anaconda (one for the hda branch and the other for the master branch), it might be worth checking if you are using the same version of Pyomo.

parkyr commented 5 months ago

Seems like the metrics of the problem are different in README.md and the results.

README: Problem vars Bool bin int cont cons nl disj disjtn
HDA Model 733 12 0 0 721 728 151 12 6

Obj: 5966.51

Results (master and hda branches): Problem vars Bool bin int cont cons nl disj disjtn
HDA Model 721 12 0 0 709 728 151 12 6

Obj: 5801.27

parkyr commented 5 months ago

Seems like the metrics of the problem are different in README.md and the results.

README: Problem vars Bool bin int cont cons nl disj disjtn HDA Model 733 12 0 0 721 728 151 12 6 Obj: 5966.51

Results (master and hda branches): Problem vars Bool bin int cont cons nl disj disjtn HDA Model 721 12 0 0 709 728 151 12 6 Obj: 5801.27

Cross-checked with the presentation and confirmed the results table values are the correct ones. Changes have been made in README.md to correct this.