A personal project that consists in the Flappy Bird game that is self-played, thanks to a Neural Network (self-coded) and Genetic algorithms.
Each iteration of the game, creates 10 birds that fly randomly through the pipes. Some of them (by pure luck) will succeed more than others, so we choose the best 3 to create a new generation of birds.
In that way, by choosing only the best birds of each generation and by repetition we achieve to adjust the Neural Network (bird brain) params in order to get a bird that successfully goes through the pipes
Main technologies:
- Java
- Java Processing
- Self-coded Neural Network (Adjustable size and nodes)
- Genetic algorithms
Captures:
