I read a fascinating paper around a decade ago that proposed, roughly speaking, to cycle through the limited set of different antibiotics that we have available in some careful order, so that as far as possible, only 1-2 are in use at any one time, ideally around the world. As soon as resistance against the incumbent antibiotic grows past a certain level, it is "benched" and replaced with the next in the sequence, with the goal being that, by the time that all other antibiotics in the sequence have been used and it's time to revisit the original one, the stresses placed on the bacterial population by the other antibiotics in the meantime have been sufficient to completely eliminate resistance to the original. This would mean that resistance to the original antibiotic the second time around would have to develop "the hard way", i.e., via novel mutations (rather than reactivation of alleles still present in the population), thus maximising the period for which it will remain effective.

Massive practical coordination problems, but I find the idea of consciously exploiting this "time dimension" really interesting.