CMA-ES
- 1 minThis library is a minimalistic Scala version of the CMA-ES algorithm as described in the The CMA Evolution Strategy: A Tutorial paper. The implementation is not as full as Java version of the algorithm, however it provides certain benefits compared to the latter:
- This implementation is more compact and straightforward.
- In Scala version objective function can be evaluated in parallel. This might help to reduce optimization time if it takes long time to compute your fitness function.
- The strategy implementation is stateless and idempotent.
Fork it here.
Vlad Orlov
Data Scientist, Open Source Contributor, Technology Enthusiast