Abstract
In Genetic Programming (GP) and most of the other
evolutionary computing approaches, the knowledge which
is learned during the evolutionary processing is
implicitly encoded in the population. In this research,
we proposed a new approach for program synthesis --
Program Evolution with Explicit Learning (PEEL), which
learns and makes use of this knowledge explicitly. PEEL
learns probability distribution from previous
generations and stochastically generates new
populations according to this distribution. On the
benchmark problems we have studied, this approach can
synthesize more compact and more accurate programs than
GP.
Users
Please
log in to take part in the discussion (add own reviews or comments).