Erstellung einer diskreten, stochastischen und adversarialen 2D-Umgebung für KI-Vorlesung

A miniature 4×5 turn-based grid world has been created where an AI robot must reach goal squares while a chasing ghost tries to catch it. Random “slip” moves and walls add uncertainty, making the Gymnasium-based environment a compact test-bed for reinforcement-learning algorithms.