Tetrapod Evolution - Genetic Neural Networks

Population size: 50 Generation: 1 - 400 Fitness function: distance * speed Selection: - Ranked - Bias 1.7 - Cutoff 0.9 - Elites: 1 Neural network: - 13 Input Neurons (8 kinestetic, 4 touch, 1 balance) - 8 Hidden Neurons - 8 Output Neurons (8 muscle) - 168 Synapses - Neuron bias range: 0 - Synapse weight range: -2, 2 - Activation function: Hyperbolic tangent approximation. Genome: - Mutation rate: 0.007 - Crossover: per bit, random - Gene count: 197 - Gene size: 2 byte - Encoding: Gray coded binary - Gene expression: synapse weight and neuron bias. - Initial state: random Body: - Tetrapod, 4 legged - Variation: Constant Sensors: - kinestetic sense (angle of joints) - balance (angle of body) - touch on the 4 legs Environment: Flat Constant