search

MDCHAMP / FreeLunch

Meta-heuristic optimisation suite for python
https://pypi.org/project/freelunch/
MIT License
46 stars 3 forks source link

How to handle submethods in the optimisers #16

Closed MDCHAMP closed 3 years ago

MDCHAMP commented 3 years ago

Some algorithms have functions that are essentially hyperparameters for example the neighbour function in SA or the mutation operations in DE and SADE.

Need to decide a way to permit users to select these functions both from a list of implemented ones and also provide their own implementations.

I think the ideal solution would be a modification to the base class that requires as little per-optimisation implementation as possible,

MDCHAMP commented 3 years ago

I think that a certain amount of abstraction for things like selection and initialisation might also be good to have at implemented at a base level.

MDCHAMP commented 3 years ago

New class system as of 1cd7b15186c940739be33f6ed6decbcbd42dedec