Genetic algorithms supporting generative design of user interfaces: Examples

Designing User Interfaces (UIs) is considered a creative and human-intensive task, and this prevents from adopting computer-aided tools to explore alternative solutions. A number of decisions are to be taken, such as optimizing the displacement of UI items on the available screen space or selecting an adequate palette of colors. Therefore, the UI design process is iterative in nature, time consuming, and costly. Generative design, while keeping the centrality of human creativity, introduces a novel approach to assist humans in finding effective and esthetically convincing solutions. Meta-heuristics are a means to apply generative design in practice. We advocate the use of meta-heuristics to supporting different aspects of the UI design process. In this paper we discuss the application of genetic algorithms to solve two different problems: building hierarchical menu layouts and selecting color palettes based on a set of constraints. The paper discusses the results of two experimental applications developed in Poste Italiane, namely the money ordering by Automated Teller Machine (ATM) interfaces and the provision of financial information by kiosks.