Can Geatpy strengthen the cooperation with Ray?

[hazeone]

Ray provides a simple, universal API for building distributed applications. Meanwhile, it has been widely used in industry. Such as Intel, Mircosoft, Antgroup... https://ray.io/ https://github.com/ray-project/ray I wish that Geatpy and Ray can make sparks in the distributed field. I can provide the necessary help in the cooperation. Wish your reply~

[geatpy-dev]

Hello! There is no conflict between Ray and Geatpy. You can utilize Ray in your aim function. Just like the way we use Scoop and MultiProcessing. Except in the case that the population size is quite big (e.g. over 1 million), the only necessary place that needs distributed calculations or parallel calculations is the aim function. i.e. The crossover, mutation and many other core operators of EA don't need to be run distributively or parallelly.

[hazeone]

I'm a developer of computing systems (streaming, distributed algorithm..) based on ray. Ray will support C++ worker this year. If you or another meet big population case need distribution support, I'm willing to support you to make this algorithm parallel. Or if the core was open source someday, I'd like to fork a version based on ray. Thanks for your reply.

[geatpy-dev]

Oh, it's very kind of you. Geatpy will be rewritten, so it's not the time to make it open source.