I can't find anything by google, but the basic approach is
probably as follows, because it makes sense: Pick a class of antibiotics that you want to remove resistance to. Stop using it. Completely. Use other antibiotics more aggressively. Selection pressure toward resisting your still-in-use antibiotics will direct evolution's optimization power toward fixing traits to resist them. If you're lucky - and I think we are, for the most part - antibiotic resistance is metabolically and/or genetically expensive, so the traits for your shelved antibiotics will be gradually lost, or even better, actively selected out. Some time down the road - I don't know how long - your shelved antibiotics will be back to full power, and you can shelve your still-in-use antibiotics out and rotate your shelved antibiotics back in.
Ideally you have enough different classes of antibiotics that you can maintain a constant rotation and always have new antibiotics coming off shelved status so you can retire antibiotics that've formed resistances. Even better, a larger library of antibiotics directly improves the effectiveness of this strategy - in the short term, introducing a new class means you can shelve an old, resisted class; in the long term, more classes means each class can spend longer on the shelf and we reduce antibiotic resistance further.