Skip to content

Conversation

@rg20
Copy link
Contributor

@rg20 rg20 commented Sep 25, 2025

Description

In branch-and-bound, to stop the search, we are currently computing the relative gap based on the working problem. When the objective offset is very large, the relative gap on the user problem can be very large even when it is small on the working problem. This PR fixes this by always computing the relative gap with respect to the user problem.

Issue

Closes #417

Checklist

  • I am familiar with the Contributing Guidelines.
  • Testing
    • New or existing tests cover these changes
    • Added tests
    • Created an issue to follow-up
    • NA
  • Documentation
    • The documentation is up to date with these changes
    • Added new documentation
    • NA

@rg20 rg20 requested a review from a team as a code owner September 25, 2025 14:00
@rg20 rg20 requested review from chris-maes and hlinsen September 25, 2025 14:00
@copy-pr-bot
Copy link

copy-pr-bot bot commented Sep 25, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@rg20 rg20 added bug Something isn't working non-breaking Introduces a non-breaking change labels Sep 25, 2025
@rg20 rg20 added this to the 25.10 milestone Sep 25, 2025
@rg20 rg20 requested review from chris-maes and nguidotti and removed request for chris-maes and hlinsen September 25, 2025 14:00
@rg20
Copy link
Contributor Author

rg20 commented Sep 25, 2025

/ok to test

@copy-pr-bot
Copy link

copy-pr-bot bot commented Sep 25, 2025

/ok to test

@rg20, there was an error processing your request: E1

See the following link for more information: https://docs.gha-runners.nvidia.com/cpr/e/1/

@rg20
Copy link
Contributor Author

rg20 commented Sep 25, 2025

/ok to test 40f19a2

Copy link
Contributor

@aliceb-nv aliceb-nv left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the super quick fix Rajesh!

@aliceb-nv
Copy link
Contributor

/ok to test 40f19a2

@rg20
Copy link
Contributor Author

rg20 commented Sep 25, 2025

/ok to test d42c7b8

@rg20
Copy link
Contributor Author

rg20 commented Sep 26, 2025

/ok to test d5edb2e

@aliceb-nv
Copy link
Contributor

/merge

@rapids-bot rapids-bot bot merged commit 96c58cb into NVIDIA:branch-25.10 Sep 26, 2025
173 of 174 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working non-breaking Introduces a non-breaking change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[BUG] cuopt incorrectly report optimal integer solution when the objective offset is very large

2 participants