I was thinking about that paper referenced in my "Where did the Universe come from? Part 3" post yesterday.
Particularly the quote:
The implication of such a description, as we have suggested in Section (1), is that Poincare recurrences are inevitable. Starting in a high entropy, “dead” configuration, if we wait long enough, a fluctuation will eventually occur in which the inflaton will wander up to the top of its potential, thus starting a cycle of inflation, re–heating, conventional cosmology and heat death.
Before I go on, I have to say that this is all idle speculation. I am not an expert on this. I may have completely misunderstood the paper. I do not believe this is the absolute, final truth. It's just idle speculation.
What this seems to be saying is that given a long enough time, the universe will naturally fluctuate from a high entropy state to a low entropy state. If this happens to the entire universe, or even just locally, say a few hundred trillion light-years in diameter (a microscopic dot compared to an infinite space), it would appear to violate 2LoT. I don't know what a Poincare recurrence is, nor do I have the time or ambition to work through the math. Maybe it doesn't.
But anyway, I like the idea of local fluctuations. Each separate fluctuation could be considered its own universe, which nicely incorporates multiple universes hypotheses. A fluctuation would be like the Big Bang, with matter and energy suddenly emerging from nothing. The physics of each universe could be randomly set at the moment of fluctuation. As a universe wound down, it would tend to spread out (expanding universe). Which could lead to interesting consequences if two (or more) of these universes were to intersect. A universe would expand until it was gone, rejoining the background nothingness. No need for expansion/contraction cycles.
This sounds like a rough basis for a science fiction setting. I may try my hand at writing again.