Evolutionary Swarm Design
Abstract
We propose swarm computing as the central paradigm that can lead to new approaches towards design. Design is seen as an unfolding process, defined through mathematical rules and algorithms, instead of static descriptions of building blocks. We introduce s warm grammars as an algorithmic aid for design tasks, where we integrate bio-inspired computational developmental models with evolutionary breeding and utilization of emergent phenomena. Swarm grammars allow intuitive, graphical and universally applicable modeling of multi-agent systems. We present evolutionary breeding techniques as examples of how to direct automated or semi-automated development of design solutions. This approach utilizes the autonomy of swarm systems through their emerging interaction processes and dares the designer to surrender total control over the design process. The computational design world, in which the designer acts like a breeder, becomes a design partner, which encourages a vivid dialogue between the swarm design system and the human designer. We further argument, that despite proclamations of the “death of theory”, the focus of design is merely shifting towards framing the respective design challenges into formalized mathematical models. Evolving and unfolding processes are accessed through simple entry points; results are promoted or demoted through interactivity on a multitude of levels. We demonstrate our first steps in the direction of such a new approach for design and discuss its consequences and challenges.