Stephen Paul King wrote:
Does computational complexity (such as NP-Completeness, etc.) and computational "power" requirements factor into the idea of simulated worlds?
It may. Also important is the issue that Tegmark raised in the Scientific American article about the ordering of an infinite set. The probability of the occurence of an element of any subset (say the even numbers) can be altered depending on how the element of the set (say the natural numbers) are ordered.
So if we assume that the multiworlds are an infinite set, to compute the probability of any event we need to know how the multiwords are ordered. I conjecture that the ordering should be anthropy related.
Let's consider a double slit diffraction experiment. The multiworlds are ordered according to the output diffraction pattern. Since the phases add up to produce this pattern, it seems that the process is linear, (thus simplifying computation) so computational complexity and computational power do seem to be of relevance.