This repo contains the code accompanying our paper,
A Comparative Analysis of Evolutionary Algorithms for Game-Playing Agent Evolution: CMA-ES vs. a Custom EA in the EVOMAN Framework
Using the Evoman framework, we test, analyse, and compare the fitness of two competing algorithms, one based on CMA-ES and one created from the ground up by us, on 3 levels of the Evoman game.
Evoman is a video game playing framework to be used as a testbed for optimization algorithms. A demo can be found here: https://www.youtube.com/watch?v=ZqaMjd1E4ZI