Directed Evolution: Evolving the Perfect Genome

Population

Population size 80

Mutation rate 12%

Selection pressure 4

Generation speed 1.0×

Evolution Stats

Generation

Diversity σ

0.00

Best fitness

Mean fitness

How it works

Each organism's genome is two real-valued traits (x, y). The terrain height is a fitness landscape — a sum of Gaussian peaks, with one tallest peak representing the "perfect life" optimum and several lower peaks as local optima:

F(x,y) = Σ h_k · exp( -((x-x_k)² + (y-y_k)²) / (2s_k²) )

Every generation runs the classic genetic-algorithm cycle:

Elitism — the single fittest organism survives unchanged.
Tournament selection — pick k random organisms (selection pressure), the fittest of the group becomes a parent. Higher k = stronger pressure, faster but riskier convergence.
Crossover — a child's trait is a random weighted blend of its two parents: child = a·parent₁ + (1-a)·parent₂, a ~ U(0,1).
Mutation — with probability = mutation rate, Gaussian noise N(0, 0.35) is added to a trait, letting the population explore off the current hill instead of freezing on a local optimum.

Watch the population climb toward the tallest peak over generations; too little mutation and small population can trap it on a shorter local hill — the same premature-convergence failure mode real directed-evolution and protein-engineering pipelines have to guard against.

Fittest organism Population

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.