General > Biology

Natural Selection as an abstraction for game theory

<< < (6/6)

Numsgil:
Wikipedia has done a better job than I have of explaining what I mean by "abstraction levels" and time frames.  Clearly I'm not the first person to examine natural selection in this way.

What I think is unique, however, is my saying that indirect fitness, that is, the fitness of relatives multiplied by their degree of relatedness, can be recursively applied to a group of increasingly non related organisms, and that each group identified in this way, can act as a unit of selection in a different context.

BTW Jez, Big Berthas did disappear in most simulations when I increased the effects of waste, particularly permanent waste.  As you say, once there was a drawback to never reproducing, we stopped seing big berthas totally dominating a simulation.

Jez:
Cool! Thanks for testing it.

Actually I originally suggested waste as a form of disadvantage for cannibots, when it comes to Berthas, was more thinking of a disadvantage for a cell getting over a certain size (rather than punishing for non repro, sterility is an advantage for some species, ie ants and if humans are anything to go by sterility is possible for individuals without affecting species). A sort of altzheimers effect similar to the waste effect I guess if body gets over a certain value.

What you've done is sensible though, you have shown that Bertha's are a program anomaly not an evolutionary one. Funny how the bots are always ready and willing to exploit any part of the program landscape, niches if you like, that haven't been given disadvantages. Does make you wonder, if we ever get the proper balances in place, how exactly the bots would mimic biological evolution.

I'll have a go at reading that wiki entry sometime soon.

Numsgil:
I don't know that it's necessarily showing that it's a program anomaly only.  It's still an interesting phenomenon where individual selection causes the eventual extinction of all life.  It hasn't happened like that in real life because the real universe has been smart enough to give every advantage an associated disadvantage

Jez:

--- Quote from: Numsgil ---to give every advantage an associated disadvantage
--- End quote ---
That's the words I was looking for!

DB has pretty well balanced disadvantages for most of the bot abilities,or should have. bots that develop to the extreme of an ability show where the disadvantages aren't balanced correctly IMO. I've never actually thought of looking at DB that way...

It is interesting, DB needs to be cruel to be even a pale imitation of the universe. Without punishing the 'weak' there is no evolution.

Numsgil:
This is a good overview of altruism in the sort of environments DB represents.

Navigation

[0] Message Index

[*] Previous page