s*r
发帖数: 2757 | 1 created by a guy in umich, to search parameter values in
high dimensions to optimize some function.
the 'fitnesss' of the estimates will depends on how well the
function got optimized. the solution with high fitness
will expand its copy number in the next generation.
quite interesting and creative. hehe, I guess biological people
will not conceive such an idea since they tries to avoid theories
incurring high genetic load. while GA quite relies on it
i guest the model includes drift, but think i |
|
