Selecting the parameters of an evolutionary algorithm for the generation of phenotypically accurate fractal patterns

This paper describes the selection of parameters of an Evolutionary Algorithm (EA) suitable for optimising the genotype of a fractal model of phenotypically realistic structures. To achieve the proposed goal an EA is implemented as a metaheuristic search tool to find the coefficients of the transformation matrices of an Iterated Function System (IFS) which then generates regular fractal patterns. Fractal patterns occur throughout nature, a striking example being the fern patterns modelled by Barnsley. Thus the algorithm is evaluated using the IFS for the fern fractal using the EA-evolved parameters.