Live, in-browser GA demos you can run without writing code.
A tour of the in-browser genetic-algorithm demos and how to run them.
A population of random vehicles tries to roll across uneven terrain. The GA evolves better cars over time.
A delivery drone finds a short tour through ten random stops on a 2D map. The GA evolves the visit order.
A genetic algorithm evolves a plan, a signpost in every cell, that guides a mouse through a fresh perfect maze from the start corner to the exit by the shortest route.
Seven antennas on a 2D map with obstacles. The GA evolves where each one sits and how far it reaches to maximise coverage and limit wasted overlap.
Rockets cross a mountain in a crosswind to land on a pad that moves every generation. The GA evolves a flight computer that lands on terrain it has never seen.
Fifteen jointed bipeds try to cross a rigid plank road full of holes and swinging wrecking balls. Every walker shares the same built-in stride; the GA evolves the body shape so the right build survives.
Steer a spacecraft to four deep-space beacons through a four-body gravity field using a fresh genetic algorithm at every decision point.
A car learns to steer around a race track using five distance sensors. The genetic algorithm evolves a tiny neural network from crashing into walls to completing clean laps.
A dino learns to play an endless runner, jumping cacti and ducking flying obstacles. The genetic algorithm evolves a tiny neural network from tripping at the first hurdle to long clean runs.
A paddle learns to keep a bouncing ball in play and smash a wall of bricks. The genetic algorithm evolves a tiny neural network from missing the ball to long rallies.
Share any demo or problem to social media with a link preview, and record a short video of a demo to post.
A bird learns to flap through a stream of pipes. The genetic algorithm evolves a tiny neural network from crashing into the first pipe to long clean runs.
A snake learns to hunt apples around a grid without boxing itself in. The genetic algorithm evolves a tiny neural network from wandering in circles to eating seventy apples a game.
An agent learns to steer across a grid dotted with blocks and patrolled by monsters to reach one gem after another. The genetic algorithm evolves a tiny neural network from wandering to collecting gem after gem.
A frog learns to dodge four lanes of traffic, then ride drifting logs across a river into one of three home bays. The genetic algorithm evolves a tiny neural network from hopping into cars to crossing again and again.
The genetic algorithm solves the classic 3 by 3 sliding-tile puzzle one move at a time, evolving a short sequence of moves before every slide.
Walk a genetic algorithm through its lifecycle one stage at a time, from the first random population to selection, crossover, mutation and the next generation.
Play Connect Four against an agent that evolves a fresh set of board-evaluation weights with a genetic algorithm before every move it makes.
The genetic algorithm repaints a target image one brushstroke at a time, searching for the stroke that brings the canvas closest to the picture.
A paddle learns to keep returning a bouncing ball against an opponent. The genetic algorithm evolves a tiny neural network from missing the first ball to long rallies.
Your feedback helps us improve. Pick a star rating; everything else is optional.