ENH: Implement crossover schemes analogous to selection schemes

uw-cmg / StructOpt_modular

A flexible and modular Structure Optimization suite for combining experimental data with energy simulations to create atomic structures.

3 stars 5 forks source link

ENH: Implement crossover schemes analogous to selection schemes #18

Closed jjmaldonis closed 8 years ago

jjmaldonis commented 8 years ago

See http://structopt-modular.readthedocs.io/en/latest/workflow.html for more info.

jjmaldonis commented 8 years ago

We should keep genetic diversity in mind.

What does it mean to have the chance of crossover be 100%? Is that all individuals mate with all other individuals, or at most len(population) children can be generated?

Should we put a cap on the number of children that can be generated? As an input parameter? That could be nice because it would be easier to define how many individuals, at most, the alg would need to relax/fit.