The multiverse hypothesis (in the style of Tegmark, et al) is partially motivated to render this sort question invalid. You're asking, effectively, "what are the hidden variables behind the multiverse", whereas the idea specifically asserts that there aren't any.
The idea is that what underlies each possible universe is just math, "all the way down".
So the question you're asking is exactly like, who created the sinusoid curve, pi, or prime numbers.
Proving the idea may be impossible. That doesn't make it wrong. If some aspect of reality is not provable, then that can't be helped.
The fascinating thing is that the hypothesis cannot be disproven either. The hypothesis that there is only one universe also isn't provable; the fact that other universes haven't been discovered/accessed doesn't prove it, since lack of evidence of existence isn't a proof of nonexistence.
The hypothesis that there are multiple universes is simpler than the hypothesis that there is just the one. It's more believable, and explains more. For instance, all remarkable coincidences in the structure of our universe are explained via the anthropic principle, which is greatly boosted by the multiple universe hypothesis.