View Single Post
Old 2020-05-16, 19:08   #5
swellman's Avatar
Jun 2012

33×109 Posts

Originally Posted by EdH View Post
Perhaps I have found the answer, in that the "raw exp_E" score range appears to be rather narrow, resulting in a large number of candidates with identical scores.
I’ve struggled with this parameter - why hang on to hundreds of likely subpar polys that seemingly take forever to complete root-opt, instead just keep the “cream of crop”. But then if one cuts nrkeep to a much lower value (say 40) has the best poly potentially been filtered out before root opt? Still 200 seems like an awful lot of interim polys to carry along.
swellman is online now   Reply With Quote