Ok how about this crazy idea:
The purpose and intent of the simulation is to develop autonomous artificial intelligences that will always behave according to a particular protocol or criteria (ie morality) so that these intelligences can then be transferred to artificial bodies in order to perform the tasks they are designed to carry out, terraforming planets, exploring environments that are hostile to biological life etc.
To this end the intelligences are subjected to a cycle of "lives" that recur until the desired stage of "enlightenment" (ie operant conditioning and/or programming) necessary has been achieved for, the isolated and highly complex tasks at hand
Our entire universe and it's metaphysical components maybe be a part of just such a production line. The programming station used to develop and input instructions into nanobots designed for terraforming or something.
Your consciousness and mine may some day be used inform the activities of a piece of nanotechnology.