• 

  A neural network

Inspiration

We were curious about genetic algorithms and their possible applications. We thought the HackUPC Fall 2016 was a good place to try to do something with them.

What it does

We applied some genetic algorithms to different visual environments to show more graphically how evolution works. Is it possible to see, in general, how the next generations of individuals are more and more prepared to overcome the challenges they face, such as approximating an image by random triangles, flying as far as possible or finishing a car race.

How we built it

We used basically html5, css3 and javascript, with an extensive use of canvas.

Challenges we ran into

Trying (quite unsuccessfully) to keep the genetic pool wide and adjusting the evolving parameters so the main goal could be reachable.

Accomplishments that we're proud of

The neural network is pretty impressive and works fine.

What we learned

Making GOOD genetic algorithms (that show diversity and converge to the main goal) is HARD HAS HELL.

Built With

• css
• html
• javascript