Reinforcement Learning target performance estimation on DR 300-5-late 19