🚀 go-pugleaf

In article <1993Jan21.172459.27732@cs.ucf.edu> long@next1.acme.ucf.edu (Richard Long) writes:

   From the few cases that I've read about the successes of GA,  
   crossover is particularly good for a problem that has a natural  
   hierarchical structure to the optimal solution, since crossover exploits  
   local fitness maxima in the construction of a global maximum.  

Not quite right.  The _building block hypothesis_ states that GAs work well
when the best solutions are _compositional_ rather than hierarchical.  By
compositional I mean that smaller pieces of the solution are pretty good but
when you get a few of the pretty good smaller pieces together then the result
is something better than all of the smaller pieces. This tops out when all of
the small pieces are together giveing a maximal fitness.

   --
   Richard Long
   Institute for Simulation and Training
   University of Central Florida
   12424 Research Parkway, Suite 300, Orlando, FL 32826
   (407)658-5026, FAX: (407)658-5059
   long@acme.ucf.edu

-pete angeline

-------------------------------------------------------------------------------
Peter J. Angeline            ! Laboratory for AI Research (LAIR)
Graduate Research Assistant  ! THE Ohio State University, Columbus, Ohio 43210
ARPA: pja@cis.ohio-state.edu ! "Nature is more ingenious than we are."

-- 
-------------------------------------------------------------------------------
Peter J. Angeline            ! Laboratory for AI Research (LAIR)
Graduate Research Assistant  ! THE Ohio State University, Columbus, Ohio 43210
ARPA: pja@cis.ohio-state.edu ! "Nature is more ingenious than we are."