On the Detection of Mutual Influences and Their Consideration in Reinforcement Learning Processes