Evolutionary design of dynamic SwarmScapes
Namrata Khemka, Scott Novakowski, Gerald Hushlak, Christian J. Jacob · 2008
This paper discusses interactive evolutionary algorithms and their application in swarm-based image generation. From an artist’s perspective, the computer-generated patterns offer departure points for creative “Imagineering. ” Input into our evolutionary system comprises of ninety-six parameters that influence the behavior of swarm agents to create emerging “SwarmScapes ” over time. This paper demonstrates how an evolutionary approach is used to create swarm-based animations, which at any point in time can be turned into electronic paintings and into high-resolution plots on canvas. As our team consists of three computer scientists and an artist, we also explore the collaborative relationships among the team members, and between the artist and the evolutionary system.