Textmeddelande: Deep reinforcement learning for real world problems