It looks like a basic chemistry may come along sooner than I thought.
In order to give the environments in Flat Galaxy an origin (and thus a way of creating themselves), I have to come up with a logical way of building their foundations. When we look at the real world, we see natural environments that have arisen all because of the interactions of basic elements--chemistry. I could build this naively and simply say things like "there are these basic elements and these particular combinations of the elements give rise to these materials and these materials make up these environments." This would be pretty simple, but it's also more arbitrary than I want. Of course, at their most basic, the laws of physics and thus chemistry are arbitrary. We just ascribe value to them because they allow our universe to exist. So far no one has discovered any fundamental reason for, say, universal gravitation to exist (and I'll go out on a limb and suggest no one ever will).
Anyway, to avoid gratuitously arbitrary rules, we'll attempt to model the underlying composition of Flat Galaxy after our own universe, but we'll avoid simulation by establishing some rules and guidelines about the basic structure.
Since we've decided that "fuel" (which represents both energy and matter, since they are the same thing anyway) should remain constant, we'll establish a starting value for that at the outset. Doesn't really matter what this number is. Everything will be scaled to it. That's the total amount of energy + mass in this universe. Now we break that number up into something. This is the starting point of the chemistry. We need rules about interactions to seed our universe with useful material. It is generally thought that the amounts of matter and anti-matter spewed out by the Big Bang should have been equal, and it's only by some crazy random stroke of luck that a fraction of a tiny fraction of imbalance between the two resulted in every bit of matter in the physical universe around us. If "random" is good enough for the real universe, it's good enough for us.
In fact, we can use randomness to our advantage all over the place.
See, instead of feeding a bunch of carefully calculated numbers into a box in order to ensure a desired outcome, why don't we just throw whatever numbers we want into the box until we get something interesting?
And how do we get something interesting? We start by presuming the existence of material (rather than throw numbers into the box until this happens), and then work towards interesting interactions between basic elements. And therein lies the rub. What are our interactions, and how do they work? We'll need a very clever solution for this.