Abstract
This paper is a preliminary report on autoconstructive
evolution, a framework for evolutionary computation in
which the machinery of reproduction and diversification
(and thereby the machinery of evolution) evolves within
the individuals of an evolving population of problem
solvers. Autoconstructive evolution is illustrated with
Pushpop, an evolving population of programs expressed
in the Push programming language. The Push programming
language can also be used in a more traditional genetic
programming framework and may have unique benefits when
so employed; the PushGP system, which uses traditional
genetic programming techniques to evolve Push programs,
is also described.
Users
Please
log in to take part in the discussion (add own reviews or comments).