Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[ISSUE] Broken comparison plots (scenario level) #9

Open
dharabor opened this issue May 19, 2023 · 5 comments
Open

[ISSUE] Broken comparison plots (scenario level) #9

dharabor opened this issue May 19, 2023 · 5 comments
Labels
bug Something isn't working

Comments

@dharabor
Copy link
Member

Algorithm comparison plots look to be buggy in some cases; e.g., orz900d-even-1. It looks like LaCAM solutions are being assigned zero cost?? See here

@bshen95
Copy link
Collaborator

bshen95 commented May 23, 2023

Don't think it is a bug. The plot compares the solution cost of each algorithm with best known solution cost found so far. In that case, LaCAM is the only one that found a feasible solution.

@dharabor
Copy link
Member Author

The plot is a mess though and doesn't communicate the situation very effectively. So I still think this is a bug.

In my opinion we should either:

  1. Track how many best-known claims have been submitted for each instance, which we could do with a stacked barplot (one coloured segment per bar for each selected algorithm being compared)
  2. We indicate the size of the gap for the best known solution and in the plot area underneath we indicate which algorithms have matched this best-known result (maybe by plotting veritcally stacked symbols)

@dharabor
Copy link
Member Author

dharabor commented May 23, 2023

Another related issue: the list of claims (in the table) tracks only how many algorithms have achieved the best-known cost. But the "Compare algorithms" plot shows dominated solutions if one exists for a selected solver. It's a bit confusing because the plot should be a summary of the table data. I think there's a place for comparing all results submitted for a specific solver vs. SOTA, on a particular map or scenario or instance, but we might want to do this as a separate feature.

@dharabor dharabor changed the title [ISSUE] Broken comparison plots [ISSUE] Broken comparison plots (scenario level) May 26, 2023
@bshen95
Copy link
Collaborator

bshen95 commented May 27, 2023

The "solution cost" function is fixed.
The "lower bound" one is waiting for trivial bound data

@dharabor
Copy link
Member Author

We seem to have trivial bounds but the plots still appear messed up. Why is there still an infinite gap? e.g.,
https://imgur.com/a/HR2oqpW

@dharabor dharabor added the bug Something isn't working label Jun 29, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants