The goal of interactive evolutionary design is to marry the exploratory skills of evolutionary computation with the aesthetic skills of the human as selective agent. A user who sees a pleasing element but who is unable to retain it in the design is likely to get frustrated with the system very quickly. The goal should be to develop a tool that is complementary, perhaps supplementary, to the designer. The designer should be in control, either using the tool to generate some initial ideas or using interactive evolution to explore the design space around an already existing idea. The designer must be the most powerful driver of the system through the design space.
In this study, we attempt to put ourselves in the mind of a skilled graphic designer, one interested in exploring a particular design space, perhaps using the tool to generate initial ideas within some user-defined boundaries. Alternatively, they might be starting with a pre-existing idea, some graphical “digital amber” (Rooke 2005; Whitelaw 2004), which they would like to perfect. They want to be able to lock “perfect” elements as they evolve (Takagi 1995), or even edit them directly. And, they would like to understand the basic principles of the system. In other words, we develop a system that is intuitive, useful and practical, and (in hindsight) adopt a number of the “ten steps to make a perfect creative evolutionary design system” (Bentley and O’Reilly 2001) resulting in a system that happens to be much simpler to understand and intuitive to use.