Integration of optimization by genetic algorithms into an L-system-based animation system

Abstract

In computer graphics L-systems represent a powerful rule-based language for modeling complex objects and their animation. However, designing objects and animations by rules is a difficult task because designers often cannot foresee the consequences of rules. This is especially true for non-experts in the domain. Therefore, we propose to enhance an L-system based animation system with evolutionary features based on genetic algorithms (GAs). These features support the designers' task of interactively modeling objects and animations. Starting from an initial population of L-system-defined objects, the computer proposes iteratively new populations based on fitness value that are determined by the designers' creative or functional criteria. Moreover, automatic optimization of L-system-defined objects/animations is possible if an appropriate fitness function can be found for a given problem. We present a concept to integrate optimization by genetic algorithms into an L-system based animation system. Typical examples, such as automatic function optimization and creative interactive design of objects, illustrate our work.