Skip to content

GRO4T/image-generation-using-one-plus-one-evolution-strategy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

47 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Image recreation using (1+1) EA

Installing dependencies

sudo apt install libsfml-dev

Fetching git submodules

git submodule update --init

Overview

In this project I've tried to implement genetic algorithm which could recreate images using only circles. At first, I've tried the classic genetic algorithm, but realised that it is badly suited for this problem. Mainly because it uses crossover, which is hard to define for images, and has static mutation rate. I've decided that (1+1) evolution strategy would yield better results.

Results

After 30min

Final thoughts

Generally I'm quite satisfied with the results. Even though, at first I planned to use multithreading I ultimately didn't do it. I felt that because algorithm only operates on single individual it would be hard to parallelize (I would have to parallelize image evaluation). If I were to use multithreading, I would probably move to (mi + lambda) evolution strategy.

Releases

No releases published

Packages

No packages published