Neural Network Evolution - Neuroevolution Simulation

Gen 0 — 0%

Fitness Over Generations

Best Creature Brain

Controls

Speed: 1.0x ?

Statistics

Generation0

Alive0

Dead0

Tick0

Avg Fitness0

Best Fitness0

Food Eaten0

Time0s

Parameters ?

World Size: 40

Population: 50

Food Supply: 80

Duration (s): 20

Hidden Neurons: 10 ?

Mutation: 12%

Crossover: 65%

Min Speed: 0 u/s

Max Speed: 6 u/s

Min Rotation: 0 deg/s

Max Rotation: 180 deg/s

Visualization

Target Vectors (Red)

Velocity Vectors (Blue)

Antenna Vectors (Yellow)

Rotation Vectors (Green)

About

Each creature has a feed-forward neural network (6 inputs, hidden layer, 2 outputs). Inputs: direction to food (sin/cos), food proximity, wall proximity (X/Y), current speed. Outputs: rotation rate and movement speed. Evolution uses tournament selection, uniform crossover, and adaptive mutation.

Neuroevolution