User Interface (UI) design remains an open, wicked, complex and multi-faceted problem, owing to the ever increasing variability of design options resulting from multiple contexts of use, i.e., various end-users, heterogeneous devices and computing platforms, as well as their varying environments. Designing multiple UIs for multiple contexts of use inevitably requires an ever growing amount of time and resources that not all organizations are able to afford. Moreover, UI design choices stand on end-users' needs elicitation, which are recognized to be difficult to evaluate precisely upfront and which require iterative design cycles. All this complex variability should be managed efficiently to maintain time and resources to an acceptable level. To address these challenges, this article proposes a variability management approach integrated into a UI rapid prototyping process, which involves the combination of Model-Driven Engineering, Software Product Lines and Interactive Genetic Algorithms.