Learning to Rank vs Reinforcement Learning in Information Retrieval - which one is preferable and why?

Question

I am trying to create an information retrieval system which can benefit from user feedback (either implicit, through e.g., click-through data) or explicit (e.g., binary feed