Towards Model-based Reinforcement Learning for Industry-near Environments