Open danielskatz opened 6 years ago
@danielskatz Can we sketch this workflow out a bit more ?
the MW part would be: a master main program that calls a number of instances of a worker program, does something with them, then calls the workers again.
Calls to workers should be scalable, as the number can change. Additionally, the number sometimes may be bigger than the number of available resources.
the GA part would be layered on top as a specific kind of MW:
we would find a GA that's written in Python, and use Parsl function to evaluate the various "genes" maybe could use something from https://github.com/handcraftsman/GeneticAlgorithmsWithPython
It would be nice to have a generic master/worker application example, and a genetic algorithm example build on top of it.