In my expereince it takes about 10 million cycles before a replicator emerges in a typical zerobot sim. Until this time, there is little need to supply zerobots with an nrg supply. One can set the costs to zero and since the bots arn't reporducing (yet) they don't use any nrg (unless a mutation causes them to shoot nrg shots or similar).
But once replication emerges, bots require nrg. Even with zero costs, some nrg must be given to the offspring and at some stage, you will want costs to provide some selection pressure. An evo sim that doesn't provide some way for these primitve bots to replace the nrg lost in reproduction won't go very far.
Just having a bunch of veggies in the sim isn't sufficient. There's a big difference between a first replicator that can barely reproduce and a replicator that can also effeciently graze veggies. I presume that many millions of cycles seperate the two in evolutionary time (I've yet to evolve a zerobot into something that can both reproduce and graze effeciently). Oh sure, if you wait long enough, some bots will mutate to shoot all the time, some of these will be firing nrg feeding shots, they will bump into veggies occasionally and the result will be a few free-floating nrg shots. Nrg transfer from autotrophs to hetertrophs will occur, but slowly, ever so slowly. Over time, reproductive nrg will be replaced. But I'm not that pateint. I want fast reproduction with lots of generations so evolution proceeds quicker. To facilitate this, I want a lot of free nrg in the sim so that reproductive losses can be replaced quickly (but not so fast that grazing provides no advantage).
Some people use autotrophs, but I want to evolve hetertrophs where there is selective pressure to forage and hunt.
Everlasting nrg shots seems like a good way to provide nrg to fledgling replicators. Chance impacts will provide enough nrg intake to allow the most primitive replicator to replenish the nrg lost in reproduction yet leave plenty of room for evolving better strategies. But how to provide the shots?
For a while, I added some simple veggies to the sim ( in addition to my 'normal' DNA disable3d veggies) that had a single gene that continiously fired nrg shots. This worked well, but eventually the veggies mutated, stopped firing and started reproducing. Being an autotroph is a tremendous advantage. The only thing stopping my veggies from taking over the sim and out competing my zerobot hetertophs was the veggy population controls. I did want veggies in the sim, as a future food source that my zerobots might eventually evolve to utilize directly, but I wanted them dumb and non-mutating, ideally with disabled DNA for maximum sim speed.
I was also looking for a feedback mechanism that would keep the nrg shot density in the sim relatively constant. Too many shots spikes the population and slows down the sim. Too few and the repoduction rate isn't what I want.
What I ended up doing is using DNA disabled veggies with a veggy feeding method that is 100% body combined with corpse mode, everlasting nrg shots and an aging cost (not too high) that kicks in after 100,000 cycles. What happens is that veggies max out their body but their only nrg intake is from random nrg shot impacts, so their nrg intake rate is the same (on average) as the hetertrophs but they can't reproduce, mutate, move or do much else. Since their DNA is disabled, the aging cost is the only cost that effects them (version 2.42.9s and beyond - in prior versions even the aging cost has no effect on bots with disalbed DNA). So, they die sometime after the aging cost kicks in. When they do die, their fat, decaying corpses supply the sim with nrg shots and new veggies get repopulated to replace them. During life, they provide grazing targets for my hetrotrophs should they ever evolve the capabilities to take advatange of them. The 100k cycle aging cost doesn't bother the hetertrophs since I want fast generation times and any zerobot >100k cycles old tends to be a big bertha anyway, hording it's nrg and not reproducing. Evolution doesn't mind if it gets killed off post 100k cycles.
Now, the neat part is the feedback loop. If the nrg desinity is high, the random nrg shot impacts supply the older veggies with enough incoming nrg to stay ahead of the aging cost for longer. Fewer veggies dying results in fewer corpses decaying and nrg density drops off. If the nrg density is too low, the aging cost kills more and faster, more veggies die, more fat corpses decaying means more nrg shots in the sim. Neat huh?
I'm still playing with the varibles - feeding rate, aging cost value and kick in point, etc. - but it seems to work pretty well. What I hope is that there will be a natural progression at some point where as soon as the hetertrophs learn to kill veggies before 100k cycles of life (they arn't getting any nrg from the sun, only from shot impacts, so even simple tactics may work) the free nrg density in the sim will naturally decline since the veggy corpses will have less body and result in fewer nrg shots than had they died later in life.
