Sovereign 8s.is mirror of standardgalactic/racer
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2024-05-15 12:47:13 -04:00
.gitignore Second week. 2024-04-14 12:24:48 -04:00
ball.py Refactor to improve performance, run trials in parallel, make code more elegant. 2024-05-13 13:17:28 -04:00
coevolve.py Fix fitness calculation, trial indices. Refactor log output, score visualization. 2024-05-15 12:15:36 -04:00
constants.py Refactor to improve performance, run trials in parallel, make code more elegant. 2024-05-13 13:17:28 -04:00
main.py Fix fitness calculation, trial indices. Refactor log output, score visualization. 2024-05-15 12:15:36 -04:00
population_manager.py Fix fitness calculation, trial indices. Refactor log output, score visualization. 2024-05-15 12:15:36 -04:00
README Add a simple README. 2024-05-15 12:47:13 -04:00
simulator.py Fix fitness calculation, trial indices. Refactor log output, score visualization. 2024-05-15 12:15:36 -04:00
visualize.py Fix fitness calculation, trial indices. Refactor log output, score visualization. 2024-05-15 12:15:36 -04:00

This is my class project for Evolutionary Robotics, taught by Josh Bongard at
UVM in 2024. The purpose of this project is just to serve as a prototype for
future research, and a demo use case for my implementation of the NEAT
algorithm.

Running main.py kicks off a series of co-evolutionary experiments that
co-evolve an environment (a topographic map of the world) and controllers for
agents (rolling balls) that are randomly placed into that world. Both the
topographies and controllers are evolved using NEAT. Several different
coevolutionary dynamics are tested, and log files, sample videos, and charts of
fitness growth are produced for each experiment.