This is my first experimentation with evolutionaty computing.
It tryies to evolve agents that draw random images, to the point these agents "learns" to draw (copy) a reference image.
- Generates random noises to an empty image.
- The noise is accepted as new matriarch if it is more aproximated to an image reference then current matriarch, or discarded otherwise.
- As a consequence of this selection, the noises tend to copy the image reference without have direct access to it.
I made a big mistake.
In my anxiety to perceive the observed behaviour faster, I optimize the algorithm to the point where the evolutionary part got removed. 🤦
The evolutionary algorithm was the most important part of this program. This program definitively was never about copying a bitmap to one location in memory to another in 3 hours.
It evolved to this point so smoothly... Didn't see it coming.
But whatever. I am working on a new versiosn using ParadiseEO & eoLib for the core part.