Current circuit: P82

[ace_dent]

Hi Miguel.

It looks like we're now crunching a bigger & more interesting circuit- any details on it?
Also I spotted from current stats (Topology page), that some circuits have higher stringent self checking (SSC), but are ranked lower than circuits with higher loose self checking (LSC). This seems odd to me. If a circuit has higher SSC, surely it is fitter than one with lower SSC, even if LSC is higher?...

Regards,
Andrew

Eg.
No.01 ~ TF:1, LSC:1.00, SSC:0.25.
No.23 ~ TF:1, LSC:0.89, SSC:0.32.

[ace_dent]

Hmmm... ok, maybe should have looked around a bit more before posting. I can see how this result can occur with the current dictionary based fitness scheme. However, should the order not be the other way round, ie. P>SSC>LSC>TF vs. currently TF>LSC>SSC>P?... but then I suppose we are more likely to get non-functioning circuits: It has great parsimony (few parts), but the Fitness is rubbish, since it doesn't do anything...

I guess finding the magic formula for efficient evolution is what makes this tricky. (OT/ I'd imagine Creationists would enjoy the idea that Evolution needs direction by some Creator? /OT).

I suppose we will just have to watch how the circuit develops over the coming days.

Cheers,
Andrew

[Scrambled Eggs]

Please see my discussion of lexicographical versus continuous evaluation functions at Evaluation Function (http://www.free-dc.org/forum/showthread.php?s=&threadid=8011). :fireboun:

[michaelgarvie]

Originally posted by ace_dent
I guess finding the magic formula for efficient evolution is what makes this tricky. (OT/ I'd imagine Creationists would enjoy the idea that Evolution needs direction by some Creator? /OT).


;) Maybe that's why evolution went so slow at the start, we had to wait a billion years until God got all the mutation rates and evolution parameters right!

:cheers: Miguel