🤖 Reinforcement Learning Simulation - 3D Model

A Blender-modelled rover learns the cliff-walking gridworld with Q-learning or SARSA above a lava canyon; tile heights, wedges and arrows show the learned values, with checks against value iteration, breadth-first search and a hidden converging agent.

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install