Skip to content

Conversation

@benvanwerkhoven
Copy link
Collaborator

This pull request replaces the previous differential evolution strategy from scipy.optimize with our own implementation. There are several reasons for this, including:

  • scipy's version does not support strings as values, which are important for tuning data types in mixed precision codes
  • our own implementation can use the search space object to perform better on constrained search spaces
  • we can more easily modify this to work with parallel runners in the future

Copy link
Collaborator

@fjwillemsen fjwillemsen left a comment

Choose a reason for hiding this comment

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

This is a nice PR that completely revamps the Differential Evolution strategy to a strategy much more suited for auto-tuning.
The tests are passing - approved.

return population


def differential_evolution(searchspace, cost_func, bounds, popsize, maxiter, F, CR, method, constraint_aware, verbose):
Copy link
Collaborator

@fjwillemsen fjwillemsen Jul 11, 2025

Choose a reason for hiding this comment

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

Officially Python variable names have to conform to this regex: [^[_a-z][a-z0-9_]*$.]

@sonarqubecloud
Copy link

@benvanwerkhoven
Copy link
Collaborator Author

Closing as this has been merged as part of #329

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants