Automatic Generation of Affine IFS and Strongly Typed Genetic Programming

Abstract

Iterated Function Systems (IFS) are recursive systems that have applications in modeling, animation and, Fractal Image Compression. IFS can be represented as arrays of floating point constants or, as shown in this paper, hierarchical data structures and therefore are good candidates for evolutionary applications. This paper describes a Strongly Typed Genetic Programming (STGP) approach to the automatic generation and evolution of 2D affine IFS, and affine maps in general. We explain the nature of the encoding used in conjunction with STGP, and demonstrate with examples its application to the inverse problem for IFS.