Consider a shop owner who has to deal with having to buy for one week from a different supplier with several different brands. Another week a brand is removed or added from the market. Yet another week, the manager decides to skip three assortments of fruit soda and exchange them with a different selection of three assortments of fruit sods, and so on and so forth.
Is dealing with such issues possible with existing public implementations of reinforcement learning?
The only research that I saw dealing with dynamically changing stuff like this is the DeepMind FTW bot playing Quake capture the flag. It deals with changing layouts to the map being played, but the implementation is not public and it doesn't resemble the inventory management situation I outlined.