search

SimonBlanke / Gradient-Free-Optimizers

Simple and reliable optimization with local, global, population-based and sequential techniques in numerical discrete search spaces.
https://simonblanke.github.io/gradient-free-optimizers-documentation
MIT License
1.21k stars 84 forks source link

Lipschitz Optimizer #27

Closed SimonBlanke closed 2 years ago

SimonBlanke commented 2 years ago

In the upcoming v1.1 I will release the Lipschitz Optimizer. . An explanation and visualization can be found on this blog.

The current source code of this new algorithm for Gradient-Free-Optimizers can be found on the v1.1 branch.

By nature of this algorithm there does not seem to be any parameters. The lipschitz-constant is automatically calculated and the default distance metric is euclidean. Maybe I will include a parameter to vary the distance metric in the future.

SimonBlanke commented 2 years ago

From the search-path you can see how the algorithm is mostly exploration at the moment:

LipschitzOptimizer