Code center > Darwinbots3

Holy Moly! Species Classification Idea that Just Hit Me.

<< < (4/4)

Numsgil:

--- Quote from: rwill128 on May 14, 2013, 04:16:46 PM ---You're right that you could implement them in the code. It's just that when you do that the entire system is subject to mutations that are likely to break the code.

--- End quote ---

Yeah, you'd need a way to disable mutations in specific bits of code for it to make sense.


--- Quote ---Maybe neural nets aren't the best tool for that, but maybe they are, I'm not sure. I just like the idea of each bot having a brain -- the framework of which is set by genetic code.

--- End quote ---

I'm just grumpy on neural nets because the gap between the coolness of its name and the lameness of the actual thing is so huge. :/

Numsgil:

--- Quote from: Botsareus on May 14, 2013, 04:28:29 PM ---But I thought the difference was that all 'neurons' on a specific given layer all fire at the same time...
Then again, they can fire one by one and then the result can be computed together. Ok, I think it makes more sense now.

--- End quote ---

Each layer executes logically simultaneously, but that doesn't mean you have to actually calculate them simultaneously.  You just need an input array separate from your output array, and you're good to go.

spike43884:
Chromosonal crossover - Well a slight improvement on the algorithm, at any reproduction event a higher rate of mutation is experienced, if the chromosones match relatively well, it'll work OKish, if they don't really match, they won't work that well...Or they just will have a rare anomaly.
Then an extra level of commands (I think we might start needing [TAB]) chromcond, chromstart and chromend which group genes into CHROMOSONES. chromosone 1 will link up to chromosone 1 in the other bot, same for 2, 3 and so on. If there is spare chromosones only half of the chromosone will actually get across (remeber, half from each parent, if 1 has an extra chromosone, good bye to half the commands). now that should truely mess up inter-species mating now.
Finally, the chromcond will always be followed by a number...so its like

--- Code: ---chromcond 0
chromstart
--- End code ---
or its like

--- Code: ---chromcond 1
chromstart
--- End code ---
these values can be input by the person who makes the bot if they put in the value, if its missing a value either 1 or 0 is implemented, 0 is a recessive chromosone, dominant chromosone being 1. these values are randomised each reproduction event. Commands can go across genes and chromosones as well (some genes actually link. yes yes they do) possibly by a more efficient system, so that combinations of chromosones cause different things (like XX, XY), if this easier chromosone command system was added chromcond wouldn't be as crucial

Navigation

[0] Message Index

[*] Previous page