Develop RL environment simulations for agent training