NK Fitness Landscape — 2D Genotype Network

This 2D companion to the 3D "Fitness Landscape: Adaptive Walk" sim swaps the continuous, orbit-controlled terrain for a genuinely different mathematical object: a discrete Kauffman NK combinatorial fitness landscape over a 6-locus genotype hypercube (64 possible genotypes), drawn as a Hasse-diagram network laid out by Hamming weight. A Wright-Fisher population still evolves by real per-locus mutation and softmax fitness-proportionate selection each generation, but here an epistasis slider K lets you dial the landscape's ruggedness as an exact, verifiable parameter — from a single smooth peak at K=0 up to a maximally rugged, many-local-optima surface at K=5, exactly reproducing Stuart Kauffman's classic (1993) result. Mini strip charts track mean population fitness, genotype diversity and the full 64-genotype fitness distribution live as you tune population size, mutation rate, selection strength and epistasis.