Joined: Dec. 2008
|Quote (midwifetoad @ Mar. 03 2009,05:46)|
|So GAs are efficient at navigating gradients, but no better than brute force at climbing towers.|
So it must piss off folks like Behe when the obvious towers are associated with things like malaria and dysentery. The skyhooks are mostly shithooks.
I used to use a GAs in my coding. I think that the descriptions here are missing one point. There are other search algorithms that will climb the hills more efficiently but they tend to get stuck on a local maxima. We tended to use variants of Newton's method.
When I read about GAs in Scientific American, I went Ahhh. GAs combines testing more of the total landscape as well as cimbing any maximas.
I may have been using GAs in a totally inappropriate way because I didn't go beyond the SA article, but it did make me a hero a few times with a C program getting an optimal answer in 30 minutes, compared to the Math graduate who's MatLab model took 3 days to calculate an answer on a model that left out most of the complexity (and the answer was wrong anway)