Step-by-step explanation:
just think about the clear disadvantages :
the key disadvantage of the LRU algorithm is that it needs a lot of book keeping on each memory access to be implemented exactly. after each access the time of the least recently used page must be tracked either using timestamps or counters. or the item/ page must be brought to the top of the stack/cache (either directly or via a link). this is highly inefficient and requires often sophisticated hardware support.
most stack or cache use cases are behaving in either FIFO (first in first out) or FILO (first in last out) style anyway.
e.g. ignoring priorities for the moment, then the rotating task queue is FIFO. and the information stack of subroutines is handled FILO.
but I would say there are enough cases, where something like this is needed.
that's why I believe you meant "real time" operating systems.