Fun with Genetic Algorithm

An unnecessary, complicated way to learn some AI

Brief

• Some carrots randomly spawn in an empty land
• They can spread nearby other carrots or randomly spawn across the map
• After a while, they rot and disappear
• 10 rabbits with 10 different genomes, they see what there's in front of their and they can rotate left/right or go forward and reverse (like cars with long rabbit ears)
• If a rabbit doesn't eat start losing life. When life is 0 they die
• If a rabbit eat a carrot restore part of his life
• When all the rabbits are dead I calculate the fitness of each one (life duration in frameCount) pick the top 4 rabbit, other 1 from the pool, brew 5 new rabbit configuration and restart
• Game ends when a rabbit can live forever

Why

• I (sadly) never worked with genetic algorithm
• I (sadly) never worked with neural network
• In general, I know very little about artificial intelligence [img]

ToDo

• WebPack
• EntityManager
• Rabbit die when they reach 0 life
• Carrot rottenizer
• RabbitManager, for creating and brewing rabbits
• Rabbit N.Net
• GeneticManager, for create a new Rabbit generation
• Low Poly 3D! (because why not?)