Hi Maciej,
Did you have a look at Bulldog yet? If so what do you think? I know I have
asked for this before and that you have been working hard on other aspects
of the simulation, but it would be really useful to be able to assign
fitness criteria on a per stick basis. That is to say, if you could specify
one stick that will contribute to the Frams health if it is kept off the
ground, say, or choose a stick to be the Frams axis and specify a preferred
direction of movement relative to that axis either for the whole Fram or on
a per stick basis. If you could assign fitness on a per stick basis then you
would have much greater control of the direction that evolution will take. I
know the that open-ended evolution of complex Frams is the ultimate goal,
but as things stand computer power is still lacking and will be for sometime
to come.
Looking forward to your reply.
Cheers,
Ander
P.S. I have said this before and I will say it again. Framstiks is f#@%&*g
brilliant. Thankyou!
Re: Bulldog
> Did you have a look at Bulldog yet? If so what do you think? I know I have
> asked for this before and that you have been working hard on other aspects
> of the simulation, but it would be really useful to be able to assign
> fitness criteria on a per stick basis.
Your bulldog is very nice, with some neurons redundant...:
(ffffffffffffffffMMMMMMMMMMMMX[1:2.550][@1:0.673][1:2.788,1:-0.985,-1:1.386][|2:1.357]FFFFFMMMMMMMMMMMMX[@0:-2.161][|-5:-0.920],,rrrrrLLLLLLLLLLLLLLLLfffffffMMMMMMMMMMMMMMMMX[|1:3.845][@!:1.000][|!:0.297][@2:2.525][|/:999.000][|-9:1.957],,FFFMMMMMMMMMMMMMMMMX[0:-1.483,-1:-715.104][@=:0.000][@!:0.182][|3:1.557]RRRRFFFFFFFFFFFFFFFFMMMMMMX[@/:-0.603,-1:0.924][/:0.026][|-9:-0.399],,LLLLLLLLLLLLLMMMMMMMMMMMMMMMMX[@-1:0.363][@G:-1.820][1:0.976][@*:-1.212][0:1.114](rrlMMMMMMMMMMMMMMX[|-6:-2.867](,,lFFFX[@2:2.966][|=:0.000]),,rrlMMMMMMMMMMMMMMX[|-9:-2.945](,,lFFFX[@-1:1.418])))
> one stick that will contribute to the Frams health if it is kept off the
> ground, say, or choose a stick to be the Frams axis and specify a preferred
> direction of movement relative to that axis either for the whole Fram or on
> a per stick basis. If you could assign fitness on a per stick basis then you
> would have much greater control of the direction that evolution will take.
Yes, it would be almost manual design ;-)
But we will try to give more abilities in design of environments
and goals.
MacKo