A project designed for Computational Intelligence course.
In this project, a simple 2D minigame is implemented.
The agent needs to learn to maneuver via neural network + evolution.
This project is an implementation of an Evolutionary Games intelligent agent using neural network and evolution algorithm. The agent is able to learn to play a variety of one-player games, by evolving a population of neural networks that adapt to the game dynamics over time.
- detailed instruction of project is available in
Evolution Project.pdf
.
Helicopter | Gravity | Thrust |
---|---|---|
![]() |
![]() |
![]() |
- Hossein Zaredar
- Matin Tavakoli
- Many thanks to Parnian Rad
- Me! (compelete some part of code as HW of my Computational Intelligence course, the init repo is available here)