Genetic Algorithm — 2D Obstacle Course Navigator

This is the 2D companion to the 3D Genetic Algorithm simulation: a completely different task solved by the same encode → select → crossover → mutate loop. Instead of evolving a target string or climbing a fitness landscape, this version evolves a real navigation controller — a chromosome of 70 steering genes decoded into a walked path across an obstacle course. Every generation, the whole population is scored by a genuine fitness function that rewards reaching the goal quickly and penalises crashing into an obstacle, the fittest individuals are picked by tournament selection, bred with single-point crossover, and mutated gene-by-gene. Click the course to add or remove obstacles and watch the population re-adapt.