Em 22-12-2009 22:23, Jan Urbański escreveu:
> o) the initial state is not really a random plan, it's usualy a
> left-deep tree (and is very inefficient) and this might skew results.
Maybe a QuickPick + SA.
http://www.springerlink.com/content/garn64gt61ju5xfa/
http://portal.acm.org/citation.cfm?doid=1559845.1559889
Att
Adriano Lange