Replies: 0 comments 2 replies
-
Sounds like a diploid vs. haploid genotype, right? It would be interesting to experiment with a diploid NEAT version, perhaps using a scheme of dominance/codominance for gene expression - although NEAT direct encoding seems to present a challenge. It would be exciting to see experiments in this area.
My understanding is that crossover technically happens during meiosis (splitting a diploid cells into haploid daughter cells), have I read this wrong? |
Beta Was this translation helpful? Give feedback.
-
Yes, I think so.
Yeah, I don't have clear ideas on how this would actually work. Any thoughts/ideas are welcome!
I'm sure you're right. I should brush up and read https://en.wikipedia.org/wiki/Meiosis before I work on this task! |
Beta Was this translation helpful? Give feedback.
-
Biological sexual reproduction works like so (high level):
This is quite different to how the canonical genetic algorithm operates, i.e. with a single 'strand' of genetic material in each individual. The biological approach is likely more powerful, as it produces more variation in the offspring, although you do have to apply rules to convert two strands of DNA into a functional phenome - so what would those rules be for something like NEAT?
Beta Was this translation helpful? Give feedback.
All reactions