Skip to content

Add several global optimization algorithms - #87

Closed
gaetanserre wants to merge 54 commits into
LeanMachineLearning:mainfrom
gaetanserre:opti
Closed

gaetanserre wants to merge 54 commits into
LeanMachineLearning:mainfrom
gaetanserre:opti

Conversation

@gaetanserre

@gaetanserre gaetanserre commented Apr 13, 2026 •

Copy link
Copy Markdown
Contributor

Implements several global optimization algorithms:

Define the evalEnv environment : a stationary environment where the reward is given by evaluating a fixed measurable function at the chosen action.

Prove that PRS is consistent.

@RemyDegenne

Copy link
Copy Markdown
Collaborator

Can you place the algorithms under Optimization/Algorithms/ ? Except for PRS, which is much more general than optimization and should go to SequentialLearning/Algorithms. Also the S in PRS is application-specific since it means search, but that algorithm can also be used for non-search applications, so another name for PRS would be better. What about randomSampling?

@gaetanserre

Copy link
Copy Markdown
Contributor Author

Indeed randomSampling is nice. I renamed PRS.

@RemyDegenne RemyDegenne left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Can you open a new PR for the EvaluationEnv file only?

Comment thread LeanMachineLearning/SequentialLearning/EvaluationEnv.lean Outdated
Comment thread LeanMachineLearning/SequentialLearning/EvaluationEnv.lean Outdated
Comment thread LeanMachineLearning/SequentialLearning/EvaluationEnv.lean Outdated
Comment thread LeanMachineLearning/SequentialLearning/EvaluationEnv.lean Outdated
@gaetanserre

Copy link
Copy Markdown
Contributor Author

See #98.

@gaetanserre gaetanserre closed this by deleting the head repository Aug 20, 2026
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.

2 participants