Evolutionary Programming and Genetic Programming

Abstract

In this chapter we review briefly two powerful evolutionary techniques; these are evolutionary programming (section 13.1) and genetic programming (section 13.2). These two techniques were developed a quarter of a century apart from each other; they aimed at different problems; they use different chromosomal representations for individuals in the population, and they put emphasis on different operators. Yet, they are very similar from our perspective of “evolution programs”: for particular tasks they aim at, they use specialized data structures (finite state machines and tree-structured computer programs) and specialized “genetic” operators. Also, both methods must control the complexity of the structure (some measure of the complexity of a finite state machine or a tree might be incorporated in the evaluation function). We discuss them in turn.

The past is present, isn’t it? It’s the future too.

Eugene O’Neill, Long Day’s Journey Into Night