Skip to content

Improve efficiency of acquisition function optimization (#954) #1275

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.

Sign up for GitHub

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

Jump to bottom
Open
wants to merge 1 commit into
base: development
Choose a base branch
from
Open

Improve efficiency of acquisition function optimization (#954) #1275

wants to merge 1 commit into from

Conversation

@JannisKst
Copy link
Collaborator

@JannisKst JannisKst commented Nov 13, 2025

Changes:

  • Added _single_local_search() to handle one local search run cleanly.
  • Introduced adaptive sigma updates over iterations.
  • Added k-exchange neighborhood generation (get_k_exchange_neighbourhood).
  • Enabled parallel execution of multiple searches using joblib.
  • Adjusted plateau-walk logic to limit retries on exhausted hyperparameters, preventing wasted evaluations.

Benchmarking:
Created a benchmark structure in SMAC3-Benchmarking to validate changes and confirm improvements in convergence behavior and runtime efficiency.

Improve efficiency of acquisition function optimization (#954)
ec13a1c 
- Optional parallelization of Local Search
- Scale standard deviation of neighborhood sampling as in Regis and Shoemaker
- Optional multiple-coordinate local search
- Limit plateau walks for exhausted Hyperparameter
@JannisKst JannisKst requested a review from benjamc November 13, 2025 05:20
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@benjamc benjamc Awaiting requested review from benjamc

Assignees

No one assigned

Labels

enhancement

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@JannisKst