Evolution Lab: Genetic Algorithm

Creatures with neural-net brains evolve to evade pursuing predators. Fitness rewards survival and eating food; getting caught is a heavy penalty. Watch evasive behavior evolve.

Each creature's genome is a neural network (8 inputs, 8 hidden, 2 outputs) that turns sensor readings into movement. Fitness = survival steps + food eaten, minus predator proximity. Predators pursue the nearest living creature. Top genomes are kept (elitism) and the rest bred by crossover with mutation.