Programmers Contest: Fit pictures on a page

Brian van den Broek bvande at po-box.mcgill.ca
Sat Jul 2 01:25:12 CEST 2005


Robert Kern said unto the world upon 01/07/2005 17:24:
> Brian van den Broek wrote:
> 
> 
>>Well, I found it ironic, but only when you add that the genetic 
>>algorithm approach came up in the context of a "best fit" problem. 
>>Survival of the fittest indeed :-)
> 
> 
> Optimization codes don't always succeed. What's the irony?
> 

Well, the punning on 'fit' (fittest in the sense of "most suitable, or 
best adapted", appropriate to Darwinian theory vs. fit in the sense of 
"to be of the right measure or proper shape and size for") amused me.

There is at least shades of irony in that the best fit in the 
Darwinian sense appropriate to genetic algorithms needn't be the best 
fit in the sense of proper size and shape. So, in suitable 
circumstances, the meaning of "best fit" in one sense being the 
opposite of what was intended in the other. (As in, my sense that it 
was ironic didn't reside solely in the genetic algorithm's possible 
failure to find the optimal solution, but how the semantic ambiguity 
of fit plugs into that in the context of the original problem.) I'd 
call that at least semi-irony.

(Yes, I get that the best fit for both evolutionary theory and the 
genetic algorithm isn't the absolutely best adapted, but best adapted 
as measured relative to the contrast class of available 
alternatives--a local maximum if you like.)

All in all, I wish I'd not hit send in the first place. This is 
perilously close to sending me into fits ;-)

Best,

Brian vdB





More information about the Python-list mailing list