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

Remove randomness by default of BBox strategies #1066

Merged
merged 1 commit into from
Oct 21, 2023
Merged

Remove randomness by default of BBox strategies #1066

merged 1 commit into from
Oct 21, 2023

Conversation

cprudhom
Copy link
Member

The main idea is to define how ties are broken in black-box strategies.
Previously, ties were broken randomly, which is a kind of admission of powerlessness.
I propose, with this PR, to make possible to declare a tie-breaker if needed or let the default one which is based on lexical ordering (=variables input order).
Note that this change has few mainly an impact on fresh start when no score are recorded, but as it drives the search to different search space, it may have a huge impact (+ or -) on efficiency.
However, now, we can tweak it.

@cprudhom cprudhom added this to the 4.11.0 milestone Oct 20, 2023
@mergify mergify bot merged commit e14ee93 into master Oct 21, 2023
7 of 10 checks passed
@ArthurGodet ArthurGodet deleted the dev_search branch October 22, 2023 16:51
@ArthurGodet ArthurGodet restored the dev_search branch October 22, 2023 16:51
@cprudhom
Copy link
Member Author

I think we should revert the PR and wait for the checks to pass.
@ArthurGodet What do you think?

@ArthurGodet
Copy link
Collaborator

Yes, I do agree that we should revert the PR until the tests have been updated.

It seems that the 5 failing tests are all based on testing the number of solutions encountered during the search (which is not surprising that they might fail with the new code that break ties).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants